I solve real-world problems at the intersection of mathematics, software, and data. After more than a decade in tech and a strong focus on applied machine learning, I now operate as a freelance machine learning engineer.

Before freelancing, I worked nearly 5 years as a full-stack data scientist in high-frequency trading, understanding how exchange networks operate at the nanosecond scale and optimizing their low-latency trading system setup. Before that, I did research on machine learning for intrusion detection systems. I also worked briefly as a C++ software engineer at the beginning of my career.

Thanks to the breadth of my background, I’m comfortable working at different levels of abstraction: from formulas to code, to architecture and infrastructure.

Over the years, I've ranked top 0.5% in worldwide machine learning competitions, gave talks at several PyData conferences, developed and led internal Python and data science training programs for business professionals, and co-authored papers on feature engineering and anomaly detection in network security.

Some interesting problems I’ve worked on:

• Research, development, and integration of AI products for the European tender market.
• Modeling the dynamics of trading systems using machine learning, probabilistic methods, and discrete-event simulations.
• Fine-tuning LLMs for prompt recovery.
• Generating adversarial inputs for text embedding models.
• MLOps for pricing services in e-commerce.
• Anomaly detection on high-throughput, real-time derivatives market data.

See the past work section for those I’m allowed to expand on publicly.