Web3 Talent

Quantitative Researcher / ML Engineer at Geneva Trading (Jan 2024 to present). Researched and deployed 3 short-horizon alpha strategies across equities and crypto futures; 2 reached live trading, achieving Sharpe ratio of 1.9 to 2.4 over a 12-month period. Maintained max drawdown under 9% via signal filtering on order flow imbalance features (vs ~18% naive baseline). ML signal generation pipeline (gradient boosting + LSTM): +14% directional accuracy over moving-average baseline on held-out tick data. Engineered 20+ microstructure features (bid-ask spread dynamics, volume profile, trade flow toxicity) as core inputs across the strategy suite. Backtesting framework with realistic slippage and latency constraints reduced live-vs-backtest PnL gap by 30%.

Before Geneva: Lead ML Engineer at Yandex (Mar 2022 to Jan 2024). Production NLP / LLM pipelines at scale: content classification serving 3M+ requests/day at 30ms P50 inference latency (distilled BERT, 6x compression with quality within 2% of full model). Generative AI deployment: YandexGPT for content markup at scale, reducing annotation costs by 97%. Feature engineering at scale on user-session signals, transferable to microstructure / time-series pipelines.

Prior: 8 ML projects at AWS ClearScale across finance, healthcare, e-commerce, and forecasting. ML credit scoring at Agora MFI (default 32% to 26%). OCR computer vision at Spectr.

London-based. Open to Remote, B2B contract via sole proprietorship. Immediate start. Comfortable across UK / EU / US business hours.