The frontier of my work
Agentic strategy development
Over the past few months I've built something bigger than any single indicator: a framework of AI agents that turn a plain-English trading idea into a finished StrategyQuant strategy — and a live oracle that loads every one into a running platform to prove it's real. You describe a strategy in a sentence; an agent writes it, and StrategyQuant itself confirms it before it ever counts.
The same approach builds the whole vocabulary — custom blocks, random groups and strategy templates — each turned from a sentence into an import-ready artifact. And it's measured, not claimed: on a set of unseen, real-world prompts, 30 out of 30 strategies loaded into StrategyQuant, single-pass, at a median of about 47 seconds each.
The full toolkit
BookAgentic Strategy Development
The flagship. Describe a trading idea in plain English and my AI agents turn it into a finished, ready-to-backtest StrategyQuant strategy — verified by loading it into a live platform before it ever counts. Measured, not claimed: 30 out of 30 unseen real-world prompts built and loaded, single-pass, at a median of about 47 seconds. I use the same framework to author custom blocks, random groups and templates from a sentence.
Custom Indicators & Blocks
Bespoke indicators, signals and AlgoWizard building blocks, coded cleanly in Java and delivered import-ready — the service behind the 135 free tools I've published on the StrategyQuant Codebase. If you can describe the edge, I can turn it into a block your Builder can actually generate with: trend, volatility, market-regime, Smart-Money-Concepts, statistical tests and more.
Strategy Validation Audit
An independent second opinion on your strategies before you risk capital. I run your databank through the same gauntlet I use for my own research: Monte Carlo stress tests, walk-forward, multi-market checks, and a statistical selection gate — the Deflated Sharpe Ratio and the Probability of Backtest Overfitting — with thresholds fixed in advance, not fitted to your results. You get a clear GO / NO-GO verdict and the reasoning behind it.
Automation & Tooling
Research infrastructure around StrategyQuant: headless sqcli pipelines that generate, retest and export without clicking, custom databank columns and ranking metrics, bug-fixed and custom Monte Carlo methods, and Python tooling for batch analysis. If your workflow has a manual step you repeat every week, it is a candidate for automation.
Machine Learning for Trading
Predictive models, feature engineering and market-regime classification with scikit-learn, XGBoost and friends — applied with the discipline to know when machine learning genuinely helps and when it is just a more elaborate way to overfit. Data pipelines, model validation and integration into your trading workflow included.
Consulting & Mentoring
Hands-on guidance for traders, quants and teams: strategy design, research process, system architecture and anti-overfitting discipline — drawn from nearly two decades in the markets and from building StrategyQuant's AI. One-off deep dives or ongoing mentoring, in English or Slovak.
Let’s work together.
Interested in working together? Fill out some info and we will be in touch shortly! We can’t wait to hear from you!