I have extensive experience in managing technology teams as well as product teams at fast moving companies. I have managed diverse teams on-site, as well as hybrid and fully remote. I believe "hybrid" works best for most technology firms. Face to face time is essential to brainstorm and build strong interpersonal relationships.
An overview of my experiences as well as technologies that I have used intensively throughout my career.
Techniques and concepts I am most familiar with:
I started learning C++ in 1998 for my dissertation on speech driven computer-aided architectural design at KU Leuven, which was under supervision of professor Herman Neuckermans and doctor Benjamin Geebelen.
Ever since, I nearly continuously used C++ as the prime language throughout my career. The language has evolved rapidly and in its C++ 20 form there are many new language concepts and syntax and I kept track of the changes.
Initially I built mostly Desktop Applications and plugins for AutoCAD.
Then, from 2001 to 2011 the focus was on Scientific Visualisation with C++ and OpenGL on UNIX and Windows operating systems.
From 2011 to 2017 I used C++ to generate powerful analytics tools to monitor financial markets, detecting market manipulation in order book driven markets as well as Request for Quote markets. The software ran on Windows, linux and Sun Solaris.
From 2018 to 2020 I worked on a 3 million line code basis at a commercial bank, improving testability, refactoring and future proofing.
From 2019 to June 2023 I built and managed a REST API for a multitenant SaaS solution on AWS to keep track of regulatory change.
7 years of working in market surveillance gave me access to a wide variety of market participants (exchanges, brokers, buyside firms, banks), asset classes as well as trading types.
Later, the opportunity to work as a Risk Manager in the Quantitative Portfolio team at Lloyds Banking Group allowed me to further refine my understanding of financial markets, and in particular Fixed Income instruments as well as Portfolio level stress testing.
Computer Graphics is how I started out coding. My journey into Computer Graphics started "properly" in 1997 by taking the Computer Graphics course in the Computer Science department at KU Leuven. I was fascinated by the boundless possibilities of CG and it bringing together skills in computer science, physics and mathematics. I enjoyed bootstrapping graphics pipelines, building up algorithms and capabilities from first principles.
Later at Flomerics and Future Facilities I worked on Computer Graphics in the context of Scientific Visualisation. Work including building particle systems, digital twins, collision detection, animation. Later still at Ancoa we made a difference in the market through the user experience of complex events in financial markets through visualisation and interactive analysis. Techniques used:
I started coding Python professionally whilst contracting at Lloyds in 2018. Python allows for rapid prototyping and is suitable for a wide range of mathematical purposes. In particular use NumPy, pandas, matplotlib, SciPy, scikit-learn and PyTorch. At Single Rulebook I extensively used Beautiful Soup, PyTorch and FastAPI to build microservices.
I have managed and collaborated closely with data scientists working on solving deep learning problems and have a sound understanding of the methodologies and techniques. I co-created workflows based on modern BERT transformer technologies in combination with more rule-based techniques such as formal language parsing. In highly regulated fields, using hybrid approaches can work well.
I am deeply familiar with the challenges around ML ops, particularly when using ML in a highly regulated environment, where we want to achieve traceability to the highest degree possible.
From a hands-on coding perspective my knowledge is particularly focused on reinforcement learning.
I am experienced with building REST APIs and microservices for distributed architectures.
This website is running on a Ubuntu linux server micro instance on AWS, using:
Whereas software engineering is ultimately just a tool, the real beauty is found in the mathematics behind it. Before diving into coding, a good engineer will look into reformulating the ask and come up with multiple solutions, using different angles of attack. Math is our toolbox to rephrase the question and come up with solutions that are both elegant and computationally efficient.
Areas of maths I am confident in: