Trading systems & AI engineering

Algorithmic trading systems built by someone who trades

Production trading infrastructure, ML-driven strategy optimisation, and risk management systems. 25 years shipping real software — from embedded C++ to live FX strategies executing right now.

Get in touch What I build
Live grid strategies — multi-pair FX
State-machine hedge management
ML regime detection in production
NL-based · BV structure · Remote-first
What I build

Trading infrastructure
& production AI

Not backtests and Jupyter notebooks — systems that manage real positions, handle real risk, and run unsupervised.

Trading systems

Algorithmic trading infrastructure

Grid-based execution engines, dynamic ATR-based spacing, real-time basket position management. Full Python stack with MT5 integration, designed for multi-pair concurrent operation.

Risk & hedging

Risk management & hedge systems

State-machine driven hedge management (diode hedges, anti-hedger recovery), position-count gating, and cooldown protocols. Built to prevent the catastrophic loss, not just optimise the profit.

ML for trading

ML-driven strategy optimisation

Regime detection models, reinforcement learning strategy selection, Kelly-criterion position sizing with regime posteriors. PyTorch, TensorFlow, MLFlow — applied to financial time series, not generic data science.

Production AI

AI/ML engineering & infrastructure

End-to-end ML pipelines, model deployment, data platform architecture. From prototype to production at scale for enterprise clients across fintech, media, logistics, and retail.

Data platforms

High-performance data systems

Real-time data pipelines, time-series infrastructure, and scalable analytics platforms. AWS, GCP, containerised deployments.

Architecture

Technical leadership & architecture

System design for complex, latency-sensitive applications. Architecture reviews, team guidance, and the hard decisions that come from decades of shipping production software.

Python C++ Go Rust PyTorch TensorFlow MLFlow MT5 / MQL5 AWS GCP Docker / K8s TimescaleDB Vector DBs LLMs FastAPI

Erik Stabij

25+ years in production software engineering — from embedded systems and real-time C++ to cloud-scale Python, ML pipelines, and now algorithmic trading systems. MSc background in physics.

I currently design and operate live algorithmic trading strategies across multiple FX pairs — grid-based systems with production-grade risk management including state-machine hedge controllers, regime-conditional sizing, and dynamic grid spacing. My edge is in position and basket management, not prediction accuracy.

On the AI/ML side, I build regime detection models, RL-based strategy selectors, and ML pipelines applied specifically to financial time series — work that sits at the intersection of quantitative finance and production ML engineering.

At StarBrain AI, I take on consulting and contract work that keeps me sharp: trading infrastructure, ML systems, and technical architecture for companies that need senior engineering depth without the overhead.

Brenntag PwC bol.com Takeaway Talpa TV de Persgroep Varian Medical
25+
Years shipping
MSc
Physics
Live
Trading systems
NL
BV structure

Let's talk

Open to contract and freelance work in trading infrastructure, ML systems, and production AI engineering. Available through StarBrain Media BV.