Background

With over 20 years of experience in software development, I have consistently worked independently on complex projects, often with minimal supervision. A key aspect of my approach is acquiring ownership of entire systems—working closely with others to gather requirements, design solutions, implement them, and ensure ongoing maintenance. Honesty, clarity, and a focus on the overall success of the project have always been central to my work, ensuring that the project’s well-being is prioritized over individual interests.

My approach to development is iterative and pragmatic. In the early stages, when a product is still evolving, or when time is of the essence, it’s often necessary to make practical decisions. However, I always emphasize the importance of maintaining code quality through periodic refactoring, keeping technical debt under control and ensuring the project remains scalable and sustainable in the long term.

In addition to my professional work, I have pursued a wide range of personal side projects across different fields, enhancing my multidisciplinary perspective. This diverse experience allows me to see the bigger picture and connect the dots between various disciplines, feeding my creative thinking and enabling me to offer well-rounded, cross-disciplinary solutions that address both technical and strategic challenges.

Consulting Approach

Effective consulting requires constant communication with the client to fully understand their situation and evolving needs. I prioritize regular discussions to assess my contribution and ensure that the solutions we develop are aligned with their goals. My main focus is on delivering ethical work and generating real value for the client.

While my collaboration can sometimes involve direct product development, my primary aim is to provide advisory support to the client’s team and design an actionable plan that aligns with their long-term vision. I offer my expertise in tackling complex problems, leveraging my strong communication skills to convey technical concepts in ways that non-technical stakeholders can easily grasp. My creative thinking, combined with a holistic, multidisciplinary perspective, allows me to bring fresh ideas to the table.

Above all, I understand the critical importance of implementing data management as an integral part of any product or business strategy, rather than treating it as an afterthought. This focus on data ensures that the solutions we implement are not only innovative but also sustainable and adaptable for future growth.

Selected case studies

Optimized Trading System and Algorithmic Analysis

In 2007, while working for an algorithmic trading firm in London, I developed a highly optimized C++ module to interface with the LIFFE futures options market, leveraging my previous experience in video game development where low-latency code was crucial. In the high-frequency trading sector, latency could make the difference between a profitable or unprofitable investment strategy. I completed this project ahead of schedule, replacing the previous, slower Java-based interface. Afterward, I had the opportunity to further improve it on my own initiative. Observing how traders developed and monitored high-frequency trading strategies, I identified the need to objectively measure algorithm performance based on executed trades and the intent behind each trade. To address this, I created a rudimentary tool that analyzed logs and generated reports, enabling earlier detection of abnormal behaviors.

Since part of my responsibilities included maintaining the operational link with LIFFE and helping detect and resolve issues with the traders’ investment algorithms, I used this tool as support in those tasks. Additionally, I developed another tool to monitor latency between various servers, helping to detect connection drops. This project was an early experience in building and using a data-driven system to support my work, providing critical insights into the performance of both algorithms and infrastructure.

Neural Networks and Data-Driven Validation

Through various personal projects involving the training of neural networks, I experienced firsthand the importance of using data to verify whether the network’s behavior generalized correctly with unseen data. This practical application reaffirmed the theoretical lessons I learned during my university years: the necessity of dividing available data into three sets—one for training, one for validating the model during development, and a third to test the model’s performance on completely unknown data.

This approach has broader implications for managing products or businesses based on data. While we may not be training neural networks with vast amounts of quantitative data, the creative use of more qualitative and diffuse data plays a crucial role in shaping a product or business. The key is validating decisions through data, and the collection of this data must follow a procedure integrated into the very use and management of the product. For example, a new product feature might not gain traction, warranting a reassessment of resource allocation, or a newly implemented policy might lead to a significant change in a critical metric, either justifying or questioning its continued use.

Music Playback Optimization Using Data-Driven Insights

In 2023, I was tasked with optimizing the response time of a music playback software system, a product for which I had developed most of the relevant code. The primary metric was straightforward: reducing the time between pressing the play button and the moment the music started to play. However, this simple goal belied the complexity of the underlying process. Files could be either local or require downloading from various sources. Metadata needed to be fetched, and factors such as the codec used, the platform on which the software was running, and the specific playback device all influenced performance.

To tackle the problem, I began by taking precise measurements of each stage in the playback startup pipeline, identifying where time was being spent and how performance varied under different playback conditions. Armed with this detailed data, I produced a comprehensive report that characterized the process beyond a single metric, allowing for a more accurate assessment. I then designed test procedures to evaluate progress and implemented optimizations, including parallelizing several time-consuming steps that had been executed in series for simplicity, as well as applying further optimizations such as improving asynchronous code and enhancing communication between threads. The result was a dramatic reduction in response time, improving performance by up to 400% in some cases. The availability of precise data was critical to the project’s success, and this data collection was permanently integrated into the product’s data strategy to monitor future changes and ensure consistent performance.