Amidst the complex ecosystem of technology and consulting, confident leaders stand out for their foresight and innovation. Holding the reigns at Janea Systems and the government-grade data store, Memurai, Mike is that beacon of leadership. Let’s dive deeper into Mike’s journey and what makes him the dynamic leader of Janea Systems. 

## A Leader Focused on Team Empowerment

As Chief Executive Officer, Mike’s role extends beyond just management; he embodies the very ethos of Janea Systems, ensuring that every stride the company takes is in sync with the evolving tech landscape. A cornerstone of his leadership philosophy? In Mike’s own words: “As leaders, our most important job is creating an environment where our team can thrive. A team that feels valued, empowered, autonomous, and connected will naturally excel, delivering value for our clients and propelling the company forward.”  

## A Tactical Trailblazer

Across two transformative decades, Mike has channeled his expertise into tech consulting services, software solutions, and the meticulous cultivation of high-performance teams. 

Mike’s trajectory at Janea Systems is symbolic of his dedication and strategic prowess. Starting as the Senior Vice President of Operations, Mike quickly ascended through the ranks, first to COO and now CEO in just over two years. His commitment to excellence and progressive leadership has seen Janea Systems surge to unprecedented heights during his tenure, growing more than 80% annually for the last two years.

## Legacy of Strategic Insight

A series of influential roles underscore Mike’s legacy of leadership. Mike’s initial foray into the consulting world began at Deloitte, where he cultivated his technology, client management, and leadership expertise. His career trajectory later carried him through a range of pivotal roles at multiple consulting firms, further honing his skills before transitioning into the software sector at Oco. During this time, he built the professional services function from the ground up and captured the attention of the place where he first started, Deloitte. The circle was completed when Deloitte successfully acquired Oco, marking Mike’s return, but this time leading the Managed Analytics services team. His dual stints at Deloitte—first as a young consultant and later via acquisition — laid the groundwork for his subsequent leadership roles, shaping a career marked by innovative thinking and strategic execution.   

Mike later joined Altisource to spearhead professional services for their newly formed Labs unit. He then took a more significant role leading business operations for the technology division, comprised of more than 1000 team members across the globe. In this role, his operational understanding came to the forefront as he streamlined processes, introduced innovative business models, and revamped the division’s financial management approach, resulting in impressive savings and revenue gains. Each chapter of Mike’s career builds upon the last, encapsulating his steadfast commitment to raising the bar in tech leadership. 

## A Comprehensive Knowledge Base

Rensselaer Polytechnic Institute’s rigorous B.S. program in Biomedical Engineering equipped Mike with an education that intersects technology with biomedical insights. This academic alchemy arms Mike with a multidimensional view, allowing him to approach tech problems with a blend of precision and creativity. 

## Outside of Work

Beyond his professional commitments, Mike treasures his personal life. Originally from Connecticut, Mike has set down roots in Natick, near Boston. Away from the tech world’s hustle and bustle, he cherishes moments with his wife, Kadie, and their three young children. 

## The Future of Janea 

Mike’s 20+ year career sits at the intersection of leadership, operations, and technology. His analytical approach, honed through leading teams across sectors such as professional services, product, and software engineering, brings a results-oriented perspective to Janea Systems. As CEO, Mike has focused intently on new market expansion, recruiting top-tier talent, and maintaining a rigorous focus on client delivery. With his strategic vision rooted in operational excellence, Janea Systems is charting an ambitious course to redefine software engineering consulting’s future boundaries.

Dive into Mike DeCerbo's transformative journey from Deloitte to Janea Systems' helm. Discover his unique leadership ethos.

Leadership Spotlight: Mike DeCerbo - Cultivating Excellence at Janea Systems

Leadership Spotlight: Mike DeCerbo – Cultivating Excellence at Janea Systems

Richard Branson says that a company is simply a group of people, and we firmly agree. So we've launched the Leadership Spotlight series to highlight the brilliant minds powering Janea Systems. Our inaugural spotlight shines on Alexis Campailla, our esteemed Founder and Chairman of the Board, who has been steering the company forward for 20 years.

## Alexis' Background

The tale of Alexis' journey begins in Italy, where his passion for coding was sparked at the tender age of nine on a Commodore 64.

![](https://www.janeasystems.com/wp-content/uploads/2023/07/image-1024x768.png)

Commodore 64

This early fascination quickly bloomed as he began crafting games in assembly language and mastering the nuances of C and graphics driver development for the IBM PC, all before he had even graduated high school.

His academic journey continued at the University of Rome, where he graduated cum laude while simultaneously making his mark in several programming competitions. This impressive track record eventually led him to be headhunted by none other than Microsoft. There, he played a pivotal role in creating the first version of the .NET Runtime, earning numerous patents and publications along the way.

The entrepreneurial bug bit Alexis next. He co-founded Wishpot, a startup focused on creating universal wish lists. As the CTO, he guided the company that was later acquired by Light in the Box.

## Alexis' Role in the Company

Fast forward to today, Alexis provides board-level guidance to three ventures; Janea Systems, Memurai, and Big Filter. Each venture represents his steadfast commitment to delivering mission-critical, enterprise-grade products. His deep involvement in these pivotal projects, combined with his innovative go-to-market strategies, is a testament to his knack for problem-solving and his crystal-clear vision for the future of technology.

## Alexis' Perspective on Leadership

For Alexis, leadership is not about having power but empowering others. He believes in fostering an environment of trust, collaboration, and respect, where every team member feels valued irrespective of their role or hierarchy. His philosophy of hiring fast learners and problem solvers intertwines with creating a culture of teamwork where engineers thrive.

## Alexis' Vision for the Future

As he looks to the future, Alexis envisions Janea Systems as a powerhouse continuing to bridge the talent-opportunity gap. He anticipates the company taking on more unconventional and challenging projects, always staying at the forefront of innovative solutions. Alexis foresees the company's evolution with a continued focus on engineering-centric solutions, ensuring a level of predictability around the end result that instills confidence in our clients.

## Advice from Alexis

Leadership, to Alexis, is just as much about character as it is about capability. He believes humility and non-judgment are vital attributes of a successful leader. His advice to those aspiring to leadership roles is to embrace the journey of continuous learning and unlearning. To maintain a work-life balance, he recommends staying grounded, investing time in activities you love, and always remembering that life, like coding, is about finding the perfect balance.

## Conclusion

Alexis Campailla’s profound technical experience and unwavering passion for engineering perfectly personify the core ethos of Janea Systems: to create an environment where engineers can thrive. As we continue learning from his remarkable journey and leadership philosophy, stay tuned for our next Leadership Spotlight post, where we will introduce you to more people shaping our companies.

Get to know Alexis, Founder and Chairman of Janea Systems. Discover his career journey, leadership philosophy, and vision for the future of an exciting portfolio of tech ventures.

Leadership Spotlight: Alexis Campailla - Janea Systems - Blog

Leadership Spotlight: Alexis Campailla - Driving Innovation at Janea Systems 

In an exciting partnership between Microsoft and Janea Systems, a whole host of iOS and Android React Native modules have been ported to react-native-windows. Adding these modules to the react-native-windows environment has solved some major technical challenges for Microsoft, its customers, and the open source community.  

It was essential that these customers could port their iOS and Android mobile applications over to a Windows environment. With these modules now available, the development teams can accelerate bringing their features and functions into a desktop setting and delight their customers across multiple platforms. 

The expertise and deep development experience of the Janea Systems team proved essential in overcoming some critical roadblocks and hurdles in producing an impressive array of react-native-windows modules.   
 

For more information on the react-native-windows ecosystem, new features, and developments, including Janea System's partnership, see the following blogspot: [Gallery App, TurboModules, and more community modules · React Native for Windows + macOS (microsoft.github.io)](https://microsoft.github.io/react-native-windows/blog/2021/03/16/64updates#%F0%9F%A4%9D-janea-systems) 

## What are modules and why are they important? 

Modules are an essential part of the cross-platform development process for building applications and features in React Native, extending the functionality, and making use of native APIs and function calls. 

The modules that were most important for Microsoft and its customers to be ported to react-native-windows are listed below: 

<table><tbody><tr><td>React-native-device-info &nbsp;</td><td>React-native-device-info gets hardware and OS-specific information. This function is crucial for optimizing app performance and enabling developers to build Windows apps that adapt to a diverse range of devices.&nbsp;</td></tr><tr><td>React-native-sensitive-info&nbsp;</td><td>React-native-sensitive-info module can get and set sensitive data. This feature allows developers to effortlessly maintain data integrity and user privacy on Windows apps, ensuring consistent and secure handling of sensitive information.&nbsp;</td></tr><tr><td>React-native-sketch-canvas &nbsp;</td><td>React-native-sketch-canvas module allows users to draw and sketch on a canvas or image. It has unblocked Microsoft clients in education by enabling pupils and teachers to incorporate digital drawing into their learning environments easily.&nbsp;</td></tr><tr><td>React-native-config &nbsp;</td><td>React-native-config module allows the addition of build-time configuration values that can be read at run time. It allows developers to manage environment-specific configurations effectively, improving app performance and consistency across various runtime environments.&nbsp;</td></tr><tr><td>React-native-permissions&nbsp;&nbsp;</td><td>React-native-permissions is used to read and request system permissions. It is vital for a streamlined user experience as it provides developers with an efficient way to request and handle system permissions, ensuring a seamless and secure app experience for end-users.&nbsp;</td></tr><tr><td>React-native-print&nbsp;&nbsp;</td><td>React-native-print is used to print PDFs. It offers developers the tools needed to simplify document handling for users by integrating printing functionality directly into their applications.&nbsp;</td></tr><tr><td>React-native-pdf&nbsp;&nbsp;</td><td>React-native-pdf is used to render PDFs. This proves critical for content-heavy applications as it allows developers to provide users with a seamless, efficient document-viewing experience.&nbsp;</td></tr><tr><td>React-native-tts&nbsp;&nbsp;</td><td>React-native-tts allows Text to Speech using system voices. It enables developers to enhance app interactivity and accessibility significantly, creating more engaging user experiences with text-to-speech functionalities.&nbsp;</td></tr><tr><td>React-native-auth0 &nbsp;</td><td>React-native-auth0 allows single sign-on to OAuth services. It provides a streamlined authentication experience for users, enabling developers to offer secure single sign-on functionalities within their apps.&nbsp;</td></tr><tr><td>React-native-gesture-handler &nbsp;</td><td>React-native-gesture-handler allows native gesture recognition on touch devices. This facilitates developers in creating more immersive and intuitive app experiences by implementing native gesture recognition in touch-enabled applications.&nbsp;</td></tr><tr><td>React-native-linear-gradient &nbsp;</td><td>React-native-linear-gradient enables developers to craft visually appealing interfaces by incorporating native linear gradients into app designs.&nbsp;</td></tr><tr><td>React-native-track-player&nbsp;&nbsp;</td><td>React-native-track-player adds media player functionality. This empowers developers to create rich multimedia applications, providing users with an integrated, high-quality media player experience.&nbsp;</td></tr><tr><td>React-native-splash-screen &nbsp;</td><td>React-native-splash-screen extends the application splash screen until dismissed. It enhances user experience during app load times by allowing developers to present visually pleasing and custom load sequences.&nbsp;</td></tr><tr><td>React-native-art&nbsp;&nbsp;</td><td>React-native-art is a vector graphics library. It inspires creativity by offering developers a toolset for integrating vector graphics into their applications, enriching the visual appeal.&nbsp;</td></tr><tr><td>React-native-progress-view &nbsp;</td><td>React-native-progress-view shows a progress bar. It is an effective tool for providing user feedback, enabling developers to display progress indicators, thus improving user experience.&nbsp;</td></tr><tr><td>React-native-screens &nbsp;</td><td>React-native-screens offers native screen handling for react-navigation. It optimizes app navigation performance, offering developers a resource-efficient tool for managing multiple screens in react-navigation.&nbsp;</td></tr></tbody></table>

## Overview of How to Build a React Native Windows Module 

This particular process of porting existing iOS and Android modules over to Windows involved a steady, incremental process, which can be summarized as follows: 

* Initial analysis - understand what the module does, its dependencies, requirements, and how it's implemented on Android and iOS. 
* Porting the module to Windows - use appropriate UWP APIs to get the same features on Windows, using good practices. Examine the code, compare the JavaScript to the native code, and examine how many classes there are. 
* Testing the module - incrementally test its features as they are developed. 
* Update documentation - add instructions on how to use the module on Windows. 
* Windows example - update the example project to also run on Windows / create a Windows example project. 
* Add tests - add automated testing / CI to test the Windows version of the module. 

## Key learnings/takeaways and what to do when developing modules in this way 

This was an excellent endeavor for both Microsoft and the Janea Systems development teams, with a total of sixteen essential React native modules being ported to Windows. From performance to accessibility, porting these modules brought React Native for Windows up to speed with other OSs. This unblocked Microsoft customers and the open source community and ensured devs could create a diverse range of applications. 

As mentioned in the process overview, it's a case of examining the existing iOS and Android react native modules and gauging the complexities of developing a react-native-windows version. It depends on the complexity of the module, how much of the code is native/C++/Java versus how much of it is in JavaScript. For example, when porting modules from mobile to desktop environments, you must be cognizant of the differences because there are cases where dependencies aren't present on desktop. On top of this, there are the added complexities of understanding the existing modules in iOS and Android, which use different languages, e.g., Swift, Objective-C, Java, etc.  

## Conclusion 

Here at Janea, we understand the importance of cross-platform development, and that's why interoperability is at the heart of many of the ground breaking projects we work on. Whether you're a developer or not, this project affects millions of Windows users worldwide. We're incredibly proud of unblocking Microsoft customers while allowing the open source community to build with, and on top of, these modules to create outstanding React Native Windows applications without compromise. 

Contact us to learn how Janea can help you with your software engineering needs [here](https://www.janeasystems.com/#lets-talk).

Discover how Janea Systems ported critical React Native modules to Windows, boosting application development across platforms.

Janea Systems Helps Unblock Key Microsoft Customers Utilizing React Native for Windows  - Janea Systems - Blog

Janea Systems Helps Unblock Key Microsoft Customers Utilizing React Native for Windows  

At Janea Systems, recruitment goes beyond just filling positions; it’s about weaving together a community of the industry’s finest minds committed to engineering excellence. Rooted in a legacy of elite software engineering, we hire individuals who excel in coding and are passionate about crafting solutions that leave a lasting impact. Our recruitment process consists of three stages designed meticulously to identify and nurture standout talent.

We understand that recruitment isn’t one-sided. Yes, interviews are a way for us to vet the best and brightest in the industry, but they’re also a way for you to see if you’d enjoy working here, too.

## Our talent selection philosophy

In the fast-paced world of software engineering, the ability to rapidly learn and adapt is paramount. At Janea Systems, we understand that the field is constantly evolving, and staying ahead means embracing continuous learning. That’s why, rather than focusing solely on specific knowledge, we prioritize the capacity for rapid learning and adaptation in our engineers. Our technical interviews are tailored to this philosophy, featuring questions that challenge candidates to solve unfamiliar problems rather than just testing their knowledge of language constructs and libraries.

Adaptability and a proactive mindset are at the heart of what we look for in our team members. We value individuals who are not only skilled problem solvers but also those who demonstrate a high level of adaptability and a can-do attitude, crucial qualities for driving our clients’ mission-critical software engineering initiatives.  
Collaboration and communication are key components of our culture at Janea Systems. We seek engineers who are passionate about their work and thrive in a collaborative environment. Our culture celebrates teamwork, encouraging each member to contribute their unique skills and perspectives.

Our philosophy ensures that our team is equipped to handle today’s challenges and prepared to lead tomorrow’s innovations. Our approach is a testament to our commitment to excellence, fostering a culture where exceptional talent thrives and innovation is a daily pursuit.

## Janea Systems Interview Rounds

Before advancing to the offer conversation and, hopefully, joining our award-winning team of elite engineers, there are three stages of interviews: 

![Janea-Recruitment-Process-Blog-1024x536.jpg](https://cloud-cube-us2.s3.amazonaws.com/nqnr0sjjo9vs/public/Janea_Recruitment_Process_Blog_1024x536_0f0925baa9.jpg)

We’ll take you through each below, explain why they’re important, and give you some tips from engineers who have been through the process themselves and made it out the other side (don’t worry, it’s not that bad).  

### **Stage 1: The Online Coding Challenge**

The Online Coding Challenge lasts around an hour and aims to gauge your ability to solve a problem using the coding language you specialize in. You’ll be tasked with coding a general solution that can handle any input within the given constraints. Authored by our engineers, it’s a pass-or-fail evaluation, with the testing platform automatically assessing your answer. 

#### Why It’s Important

This stage is crucial as it lays the groundwork for evaluating your technical skills and problem-solving ability within your area of expertise. 

#### Tips for Acing the Online Coding Challenge

A thorough understanding of coding principles within your chosen language and getting accustomed to common coding challenges will steer you through this stage effectively. These [preparation tests](https://www.hackerrank.com/dashboard) are a good place to start. 

### Stage 2: Recruiter Interview

At this stage, you’ll be invited to a conversation, lasting no longer than an hour, with a member of our recruitment team to delve into your previous experiences. This interview also provides an opportunity to learn more about who we are, what we stand for, and how we work as a company. 

#### Why It’s Important

This conversation is a two-way street. We get to learn more about you, and you get a glimpse into Janea Systems’ culture, laying the groundwork for a potentially great fit. 

#### Preparing for Success

It should go without saying, but read the job description. Try to align your experience with the job description. Also, having a clear narrative about your professional experience and a couple of questions about Janea Systems can lead to a better discussion.  

### Stage 3: Technical Interview

You’ve made it through the Online Coding Challenge, so you’ve shown some programming and problem-solving competency. You’ve made it through the Recruiter Interview, so we think you’re a good fit for the company. The final hurdle is a 2-hour Technical Interview, a deep dive into your technical and collaborative abilities. 

The 2-hour lab coding interview is designed to be dynamic and engaging. We start with a couple of logic problems. These are not just warm-up exercises but also an opportunity for us to understand your approach to problem-solving and logical thinking. 

Following the initial logic problems, we dive into a live coding challenge. As per our commitment to flexibility, you can choose any programming language you are comfortable with for this part of the interview. This approach allows us to see you in your element, using the tools you know best, which is crucial for a fair and effective assessment. The technical interview gives us insight into not only your coding skills but also how you approach problems and iterate on your solutions. It’s also a chance for us to see how you work under a bit of pressure and how you manage time and resources. 

Great engineering is as much about working well with others as it is about technical expertise. We’ll be looking at how you communicate your ideas, respond to feedback, and adapt to new information or challenges. 

#### Why It’s Important

This intensive evaluation goes beyond mere coding. It embodies Janea Systems’ philosophy of holistic problem-solving paired with robust collaboration. Our approach to software engineering isn’t just about code but revolves around critical thinking, effective collaboration, and delivering impactful solutions. 

#### Tips from our Technical Interviewers

1.  **Don’t Rely on IDEs:** You won’t have access to an Integrated Development Environment (IDE). Expect to code without the aid of auto-suggestions or syntax help. This format aims to gauge your ability to think and code without external aids. 
2.  **We’re Looking for Expressive Coding:** Our primary focus isn’t to test your memory of syntax. We’re keen on seeing how you translate a coherent thought process into functional code. The objective isn’t specific knowledge but an expressive representation of your solution. 
3.  **We’re Assessing Critical Thinking:** It’s not just about coding. It’s about aligning your solution with the company’s goals, project requirements, and client needs. Demonstrating a capacity to evaluate situations and pivot your solutions accordingly critically is paramount. 
4.  **Understand the Problem Before Coding:** A common pitfall many candidates stumble into is jumping directly into coding. Taking a moment to comprehend the problem fully can be the difference between a practical solution and a convoluted one. 
5.  **Respond to Feedback**: During the interview, you may receive feedback. Ignoring it can be detrimental. If the interviewer offers a direction or perspective, it’s worth considering. 
6.  **Simplicity Over Complexity:** A crucial tip to remember is the essence of simplicity. Given the time constraints, our problems are designed to be succinct. Overcomplicating solutions or attempting to draft extensive code might not be the best strategy (we’re not expecting 500 lines of code in 2 hours). 
7.  **Get a good night’s sleep before the interview.**

Displaying a balanced blend of technical knowledge, clear communication, and a pragmatic approach to problem-solving will ensure you’re on the way to joining the ranks of Janea Systems’ software engineering team. 

Understanding that each engineer has unique strengths, we try to be as flexible in the interview process as we are in our work. Choices, such as allowing you to pick which programming language in your technical interview, are not just about accommodating your preferences; it’s about seeing you at your best and most confident. It’s an opportunity for us to see how you apply your skills to an unfamiliar challenge in a familiar context. We believe this is crucial for a genuine assessment of your capabilities.

Adapting to various challenges and environments indicates how you will thrive while working on our diverse and ever-evolving projects. It’s these qualities that define our team at Janea Systems and enable us to consistently deliver excellence on projects that impact millions of developers and billions of people every day.

## Resources to Help You Prepare

[Daily Coding Problem](https://www.dailycodingproblem.com/) 

[Practice for Cracking Any Coding Interview](https://www.geeksforgeeks.org/practice-for-cracking-any-coding-interview/)

[The Myth and Reality of the 10X Engineer](https://www.janeasystems.com/blog/the-myth-and-reality-of-the-10x-engineer/)

## Conclusion

At Janea Systems, we’re thrilled to have brilliant minds like yours contribute to our mission of delivering unparalleled software solutions. If you’re ready to embark on a fulfilling career in software engineering, [check out our open positions here.](https://www.janeasystems.com/current-openings/)

Good luck in your interviews!

Aspiring to join our team of elite software engineers? This blog details each step of our recruitment process, offering tips and insights to prepare you for success.

The Janea Systems Recruitment Process - Blog

The Janea Systems Recruitment Process

PyTorch, hailed as the go-to open-source machine learning library, is trusted worldwide by developers and companies such as OpenAI. Initially built for Linux, Microsoft has entrusted Janea Systems with the upkeep of the library on Windows since 2022, bolstering our commitment to democratizing access to machine learning and ensuring PyTorch remains an inclusive and robust platform for Windows users.

There was an open issue on the Pytorch Github where a contributor noticed that matrix multiplication was over five times slower on Windows. With matrix multiplication being a fundamental operation within machine learning, we stepped in to help. Read on to learn about matrix multiplication and how our Senior Software Engineer, Ionut Manta, investigated the problem.

## What is Matrix Multiplication: The Heart of Computational Tasks

Matrix multiplication is a fundamental operation in mathematics and computer science. It's used extensively in various fields, such as physics, engineering, computer graphics, and machine learning. At its core, multiplying matrices involves combining two bi-dimensional arrays of numbers (matrices) to produce a new matrix. This operation is crucial for performing linear transformations, solving systems of linear equations, and manipulating data in algorithms.

Matrices represent datasets, weights, and transformations in neural networks. Multiplying matrices helps in forward propagation and backpropagation algorithms, essentially enabling the training and functioning of neural networks.

## Benchmarking Matrix Multiplication Operations: Linux vs Windows

Benchmarking the time taken to complete a matrix multiplication operation between Windows and Linux, PyTorch contributor [linssswww](https://github.com/pytorch/pytorch/issues/26345) found that when the size of the input matrix is large, their example used a matrix with 1,000,000 rows and 4 columns ((1000000, 4)), the time taken to complete the operation on both platforms is similar (around 0.101 to 0.109 seconds respectively). However, when the size of the input matrix is small, 1,000 rows and 4 columns ((1000, 4)), the performance on Linux is roughly 5.5 times faster than on Windows (0.031242s vs 0.00567626953125s)

### Impact on Small Matrix Operations

When a machine learning application involves numerous operations on small matrices, even the slightest delays add up, leading to significantly longer processing times. For instance, if a machine learning algorithm needs to perform thousands of these small matrix multiplications as part of its computation—such as during the training phase or when making predictions—the additional time taken for each multiplication can make for a substantial delay.

### Examples of the Impact of Slow Matrix Multiplication Operations in Machine Learning

Consider a real-time machine learning application, like a recommendation system on an e-commerce website that relies on matrix operations to predict and display product recommendations as a user browses instantly. This system might use small matrices to represent user and product features, and through matrix multiplication, it computes the likelihood of a user being interested in a particular product.

On Windows, where the operation takes approximately 0.031242 seconds for a small matrix, as opposed to 0.00567627 seconds on Linux, the system would experience latency issues. This difference might be negligible for a single recommendation, but the latency adds up in scenarios where thousands of recommendations are generated simultaneously. This could result in a slower response time, potentially impacting user experience by making the website appear sluggish.

### Broader Implications

In a broader scope, developers who are running experiments and training models on Windows machines might find that their iteration time is slower due to these performance issues. This can extend the time needed for model development and experimentation. For machine learning engineers who need to quickly train and test multiple models to find the best one, a slower iteration cycle could lead to longer project timelines and increased costs, especially if they are relying on cloud-based Windows machines where compute time is billed by the minute or second.

## Investigation and Technical Insight

Ionut Manta, Senior Software Engineer at Janea Systems, looked into reproducing and investigating the issue, boiling it down to [a small example](https://gist.github.com/mantaionut/e4c588e2f5273f75e8eb286a54f20c69). His work revealed that the performance discrepancy stemmed from the initialization phase of matrix multiplication operations. While the first call to the function exhibited similar timings across both platforms, subsequent calls on Windows took almost twice as long. This behavior was particularly noticeable in small matrices operations, where initialization time became a significant performance factor.

Further exploration traced the issue back to specific code segments within PyTorch's CUDA implementation. By creating a minimal example that isolated the initialization phase, Ionut not only confirmed the bug but also provided a clear path for addressing it. From here, Ionut reported the issue to NVIDIA, contributing to a broader understanding of the performance challenges across platforms.

This work started with analyzing the full PyTorch codebase, taking us through loops of benchmarking subsections of the code path that multiplies matrices. Ionut investigated several sections of the code, such as memory allocation, performance standard containers, and optimizations done by different compilers. By excluding sections with similar performance, Ionut was able to boil the issue down to the CUDA section. One major difficulty working on this was that the performance difference did not show on the first invocation, so when the code was reduced to its simplest form the difference disappeared.

## The Benefit of Matrix Multiplication Performance Enhancement on PyTorch

The issue investigated by Ionut and the team has significant implications:

*   Efficiency: For applications that frequently perform small matrix multiplications, reducing the initialization time on Windows platforms means computations can be done faster, improving the overall performance.
*   Resource Optimization: More efficient matrix operations can lead to lower CPU and GPU usage, saving energy and allowing for more computations to be done in parallel.
*   User Experience: Developers and researchers using PyTorch on Windows for machine learning tasks can expect smoother, more consistent performance, especially in applications requiring rapid computations, such as real-time data analysis or interactive simulations.
*   Accessibility: By improving performance on Windows, PyTorch becomes more accessible to a broader audience, many of whom use Windows as their primary development environment. This can help democratize access to machine learning tools and resources.

## Conclusion

The initiative to diagnose the matrix multiplication performance issue on Windows platforms represents a significant step forward in ensuring PyTorch's reliability and efficiency across operating systems. Improving PyTorch not only enhances the user experience for developers and researchers but also strengthens the foundation of PyTorch as a versatile tool for machine learning and deep learning applications. As we progress, the resolve and collaboration demonstrated by the PyTorch community will continue to drive innovations and optimizations, making advanced machine learning libraries like PyTorch tools more accessible and efficient for everyone.

**Ionut Manta is a Senior Software Engineer at Janea Systems. Learn more about him** [**here**](https://www.janeasystems.com/team/i-manta)**.**

**This particular issue is one of hundreds we've looked into on PyTorch. Find out more about our ongoing work ensuring PyTorch performs optimally on Windows** [**here**](https://www.janeasystems.com/client-success/pytorch-for-windows)**.**

Explore how we diagnosed PyTorch matrix multiplication performance issues, leading to valuable feedback for NVIDIA's CUDA team for enhanced ML performance.

Janea Systems Diagnoses Major Matrix Multiplication Operation Performance Issue on PyTorch 

PyTorch is one of the most popular open source deep learning frameworks alongside Tensorflow. With a large active community and various third-party tools built on top of it, the framework is used by the likes of Meta, Uber, and Microsoft, and it’s been widely adopted in academic research and mobile/edge deployments.

For the past two years, Janea Systems has been maintaining PyTorch on Windows on behalf of Microsoft, ensuring parity across different OSs so that everyone can benefit from PyTorch. In this article, we’ll explain how we ensured a crucial module worked on PyTorch for Windows, enabling seamless sparse matrix operations and expanding the framework's capabilities for Windows developers.

## What are Math Kernel Library (MKL) and Sparse Matrices?

MKL (Math Kernel Library) is a software library developed by Intel that provides optimized mathematical functions for various computational tasks.

The MKL Sparse Matrix module is a part of this library that specifically deals with sparse matrices. Sparse matrices are special types of matrices that contain mostly zero values, with only a few non-zero entries. They are commonly used in scientific computing and data analysis to represent large datasets efficiently.

### The benefits of MKL Sparse:

Efficient storage: It stores only the non-zero values of a matrix, saving memory and computational resources.

Fast operations: It provides optimized algorithms for performing mathematical operations on sparse matrices, such as addition, multiplication, and solving linear equations.

Performance optimization: By using specialized techniques for sparse matrices, MKL can significantly speed up calculations compared to treating them as regular dense matrices.

Integration with other tools: It can enhance the performance of various programming languages and scientific computing environments.

### Where Are Sparse Matrix Computations Used?

Sparse matrix computations are widely used in various fields, particularly where large datasets with many zero elements are involved. Common applications include:

1.  **Scientific Computing and Simulations:** In physics and engineering, sparse matrices are used to model complex systems with many variables but few interactions. For instance, simulating the structural properties of materials or the dynamics of fluid flows.
2.  **Graph Algorithms:** In social network analysis, sparse matrices represent connections between users. In transportation networks, they model the routes and connections between different nodes, such as cities or intersections.
3.  **Machine Learning and AI:** In natural language processing, sparse matrices are used to represent text data where most word pairs do not co-occur. In recommendation systems, they model user-item interactions where only a few items are relevant to each user.
4.  **Image and Signal Processing:** Sparse representations in image compression algorithms like JPEG reduce file sizes by only storing significant pixel values. In signal processing, they are used to efficiently extract important features from large datasets.
5.  **Embedded Systems and AI on the Edge:** Resource Optimization and Real-Time Processing: Sparse matrix computations reduce memory usage and computational load, making them ideal for devices with limited resources. For example, AI model compression uses sparse matrices to minimize the size and increase the efficiency of models deployed on edge devices, such as smartphones or IoT devices. Sparse matrices also enable fast computations necessary for real-time applications. Autonomous vehicles use sparse matrices for real-time object detection and path planning, ensuring quick and efficient decision-making.

## Bringing MKL Sparse to PyTorch on Windows

MKL Sparse wasn't supported on windows, leading to problems using some functions. Specifically, ‘mkl\_sparse\_d\_create\_csr’ and ‘mkl\_sparse\_destroy’ use ‘malloc/new’ and ‘free/delete’ for memory management, which requires the Universal C Runtime (UCRT). PyTorch links against the static variant of MKL with the ‘/MD’ option, meaning it uses a version of the runtime library designed for dynamic linking. However, with Microsoft's Visual Studio Compiler (MSVC), you cannot mix different runtime library options (/MD and /MT) in the same program. This conflict resulted in errors and prevented sparse matrix operations from working correctly. In simpler terms, you can't have two different versions of the same foundational software trying to work together, which caused problems.

## Solution

We ensured that both PyTorch and MKL use the same C runtime library (/MD), avoiding conflicts that previously led to linker errors. Specifically, we resolved the "LNK2038: mismatch detected for 'RuntimeLibrary'" error, where the value 'MT\_StaticRelease' didn't match the value 'MD\_DynamicRelease'. This consistency is crucial for stability and successful compilation.

Consistent use of the same C runtime library (/MD) for PyTorch and MKL resolves linker errors and streamlines the build process, making it easier to compile and maintain projects on Windows.

By removing conditional compilation for Windows-specific cases related to sparse operations, our pull request makes the code more maintainable and reduces the need for platform-specific workarounds, enhancing the overall development workflow and cross-platform consistency.

## Conclusion

We’re incredibly proud of our continued work maintaining PyTorch on Windows. By enabling sparse matrix operations, this fix unlocks new potential, from self-driving cars to advanced physics, enhancing computational efficiency and performance. As we continue to push the boundaries of software engineering and data science, such advancements are crucial in driving innovation and solving complex problems. The journey of integrating MKL sparse operations into PyTorch on Windows underscores the collaborative spirit of the open-source community and the relentless pursuit of excellence in software development.

Take a look at the original issue and the fix on GitHub here:

[https://github.com/pytorch/pytorch/issues/97352](https://github.com/pytorch/pytorch/issues/97352)

[https://github.com/pytorch/pytorch/pull/102604](https://github.com/pytorch/pytorch/pull/102604)

Ioan Manta is a Senior Software Engineer at Janea Systems. Learn more about him [here](https://www.janeasystems.com/team/i-manta).

Discover how Janea Systems enabled seamless sparse matrix operations on PyTorch for Windows. Learn about the benefits for developers, data scientists, and researchers. 

Janea Systems enables the MKL Sparse Matrix Module on PyTorch for Windows 

Janea Systems Enables the MKL Sparse Matrix Module on PyTorch for Windows 

# Janea Systems is Heading to NVIDIA GTC 2025 

We’re packing our bags for [NVIDIA GTC 2025](https://www.nvidia.com/gtc/), happening on **March 17-21 in** **San Jose, CA****.** It’s one of the most prominent AI events of the year – a place where innovation meets real-world application. As experts in AI/ML software solutions, we’re ready to connect, learn, and share insights on how to make AI work smarter, faster, and better. 

## About NVIDIA GTC 

NVIDIA’s GPU Technology Conference (GTC) is where the best minds in AI come together. Every year, tens of thousands of attendees — developers, engineers, business leaders, and researchers — gather to explore the latest in AI, deep learning, and accelerated computing. The event features a [keynote from NVIDIA CEO Jensen Huang](https://www.nvidia.com/gtc/keynote/?regcode=no-ncid&ncid=no-ncid), over 1,000 expert-led sessions, 400+ interactive exhibits, technical workshops, and exclusive networking opportunities. 

This year’s GTC is packed with insights into: 

*   Generative AI & large language models (LLMs) 
*   AI infrastructure & MLOps 
*   Autonomous systems & robotics 
*   Data science & optimization 
*   Digital twins & simulation 
*   Edge computing & IoT AI 
*   Healthcare & life sciences AI 

With so much to explore, GTC 2025 is the ultimate opportunity to engage with experts in AI development, deployment, and optimization. 

## Meet Janea Systems team on NVIDIA GTC 

For over 20 years, Janea Systems has provided enterprise-grade software solutions, helping organizations navigate their AI/ML maturity journeys. Here’s what we do: 

1.  **Align AI initiatives with business goals.** Our experts identify high-value AI use cases, maximize ROI, and minimize deployment risks. From concept to implementation, we ensure AI solutions drive measurable impact. 
2.  **Refine and structure data for optimal AI performance.** Our data engineers develop robust data pipelines, ensuring structured, high-quality, and accessible datasets. 
3.  **Develop Edge AI solutions for real-time intelligence.** Our team creates AI-driven applications that run on edge devices, enabling instant decision-making without relying on cloud connectivity. 
4.  **Optimize AI infrastructure.** We enhance scalability, reduce latency, and improve computational efficiency to ensure seamless AI model deployment on cloud, hybrid, and on-premises environments. 

Whether you’re looking to scale AI initiatives, fine-tune your models, or discuss the latest industry trends, we’d love to chat. 

We invite you to meet our team during the conference: 

[Toby Rawlinson,](https://www.linkedin.com/in/tobyrawlinson/) Chief Business Officer: With a rich background in engineering consulting and business development, Toby has collaborated with industry leaders like Microsoft and Goldman Sachs. 

[Mario Stalder,](https://www.linkedin.com/in/marioastalder/) Business Development Executive: With over 7 years of experience in business development and M&A within private equity and the tech startup ecosystem, Mario excels in building key business relationships and leading outbound sales initiatives. 

[Mirza Baig](https://www.linkedin.com/in/mrbaig/), Business Development Director: Mirza has held leadership positions at companies like Softvision (acquired by Cognizant) and Microsoft, where he spearheaded digital transformation initiatives and delivered modern enterprise IT solutions. 

[Bill Sanders](https://www.linkedin.com/in/billsanders4/), Business Development Director: After 25 years at Microsoft, Bill transitioned into tech consulting, where he has spent the last 7+ years helping businesses implement AI and cloud technologies. 

[Benedetto Proietti,](https://www.linkedin.com/in/benedettoproiettimathematician/) Senior Software Architect: With a strong background in low-level programming and optimization, he has worked on projects ranging from porting the GCC compiler for the APEmille SIMD supercomputer to contributing to massive transactional systems at Expedia. More recently, Benedetto has been involved in developing in-memory databases compatible with MongoDB and enhancing 3D tiling pipelines at Cesium. 

Don't miss this opportunity to connect with us at **NVIDIA GTC 2025!** Reach out to us on LinkedIn or simply book your meeting via the contact form below. 

See you at GTC!

We're going to NVIDIA GTC 2025 AI conference to learn about new advancements in AI and robotics and network with industry experts.

Janea Systems is Heading to NVIDIA GTC 2025

**How do you fight financial crime with synthetic datasets? Can AI support regulatory compliance? What role can agentic AI play in M&A processes?**

These are just a few of the questions we’ll be exploring in New York on **April 15–16**, as Janea Systems takes part in the **RE•WORK AI in Finance Summit** - a premier event bringing together experts in machine learning, AI, and financial services. 

With years of experience building real-time analytics systems and robust ML infrastructure, we’re excited to contribute our voice to the dialogue - and to learn from others driving AI innovation in finance. 

## About RE•WORK AI in Finance Summit

The **RE•WORK AI in Finance Summit** is where the brightest minds in financial services and AI meet to share real-world use cases, challenge the status quo, and chart the future of intelligent finance. Attendees will: 

**Learn from cutting-edge AI applications** in fraud prevention, conversational AI, and synthetic data generation. 

**Connect with industry leaders** from organizations like HSBC, Vanguard, NASDAQ, Wells Fargo, Fidelity, and Capital One. 

**Access exclusive content** - from deep-dive talks and panels to downloadable post-event reports. 

**Key themes include:** 

*   Generative AI 
*   Deep Learning and Neural Networks 
*   Machine Learning for Fraud Detection 
*   NLP for Chatbots 
*   AI Ethics and Regulations 
*   MLOps 
*   Synthetic Data 
*   Optimizing ML Models

## Meet Janea Systems at RE•WORK 

For over 20 years, enterprise leaders have trusted **Janea Systems** to deliver mission-critical software with unmatched precision and quality. 

Clients rely on us to tackle complex challenges at the intersection of AI/ML, [middleware](https://www.janeasystems.com/blog/middleware-core-banking-modernization), systems engineering, and financial technology. 

Here are just a few ways we’re making that real: 

**Building a future-proof AI-ready data lake** for a collections platform, enabling historical data tracking and machine learning analytics across fintech portfolios​. 

**Rebuilding critical fintech middleware** from the ground up, removing vendor lock-in, and enabling seamless integration with 9 banking cores. 

**Developing an exploratory AI fact-checking prototype**, combining LLMs with structured pipeline architecture for fast, cost-effective experimentation​. 

**Optimizing LLM-based scouting tools** for NFL analytics, using advanced caching, classification, and modularization to reduce latency and cost​. 

Whether it’s rethinking financial intelligence, accelerating time-to-market, or validating AI-powered proof of concepts, we engineer AI systems that can perform under pressure. 

## About Mario Stalder 

[Mario Stalder](https://www.linkedin.com/in/marioastalder/), Growth Director at Janea Systems, will represent the company at the event.

With over seven years of experience in business development and M&A within private equity and the tech startup ecosystem, Mario brings a sharp understanding of both the technical and business aspects of AI initiatives. He specializes in building strong client relationships and leading strategic outreach that aligns with long-term innovation goals. 

## Let’s Connect

Don’t miss the opportunity to connect with us at [RE•WORK AI in Finance Summit](https://ny-ai-finance.re-work.co/)! Let’s chat over coffee, discuss your organization’s AI challenges, or explore ways we can collaborate. 

📅 **April 15–16**   
📍 **New York City** 

**See you in NYC!**

Join Janea Systems at RE•WORK AI in Finance Summit in NYC, April 15–16, to explore AI in finance, synthetic data, fraud detection, and regulatory compliance.

Janea Systems at RE•WORK AI in Finance Summit NYC 2025

AI, Finance, & NYC - Janea Systems Joins the Conversation 

Janea 

large_Janea Systems OG Image No Text.png

medium_Janea Systems OG Image No Text.png

small_Janea Systems OG Image No Text.png

thumbnail_Janea Systems OG Image No Text.png

Janea Systems OG Image No Text.png

Janea Systems is a software engineering consultancy, serving the enterprises that drive global innovation for over two decades. We help our clients build and optimize technologies used by millions of developers and billions of consumers. 

Related Blogs

Let's talk about your project

Let's talk about your project