Building AI-driven quantitative research and algorithmic trading systems
Market microstructure · Multi-factor alpha research · Mid-frequency trading
Security researcher · Advocate for decentralized, self-custodial and private systems
AI-driven, multi-agent research infrastructure focused on discovering, validating and executing systematic trading strategies on Hyperliquid.
Combining market microstructure analysis, multi-factor alpha modeling, cross-exchange intelligence and machine learning to investigate predictive signals across mid-frequency trading horizons.
flowchart TD
A[Multi-Source Market Data] --> B[Data Engineering & Feature Extraction]
B --> C[Market Microstructure & Order Flow]
B --> D[Derivatives & Wallet Intelligence]
B --> E[Cross-Exchange & Macro Signals]
C --> F[Multi-Factor Alpha Modeling]
D --> F
E --> F
F --> G[AI Research Agents & ML Models]
G --> H[Signal Generation & Regime Detection]
H --> I[Walk-Forward & OOS Validation]
I --> J{Statistical & Economic Validation}
J -- Fail --> G
J -- Pass --> K[Cost-Aware Strategy Selection]
K --> L[Hyperliquid Execution & Risk Management]
Note
Predictability ≠ Profitability. Research focuses on statistically robust, out-of-sample signals that remain economically viable after fees, spread, slippage, funding and execution costs.
Research Areas & Capabilities
- Level 2 order book analysis, liquidity depth and spread dynamics
- Order flow imbalance, trade aggressor analysis and price impact
- Market liquidity, volatility regimes and microstructure signals
- Open interest, funding rates and perpetual futures basis
- Liquidation events, liquidation proximity and cascade dynamics
- Wallet fills, position tracking and large trader behavior
- Smart money analysis, wallet consensus and HLP activity
- TWAP execution patterns and large trade detection
- Cross-exchange price discovery and lead-lag relationships
- Binance, Bybit, OKX and Deribit market intelligence
- Options, equity reference data and macroeconomic indicators
- Hyperliquid HIP-3 equity-linked perpetual markets
- Multi-factor alpha research and predictive modeling
- LLM-driven autonomous research agents
- Time-series foundation models (e.g. Kronos)
- Feature engineering, regime detection and signal fusion
- Statistical validation, walk-forward testing and backtest analysis
- Transaction cost modeling and execution-aware optimization
- Mathematical invariants and property-based testing
- ~228 GB Hyperliquid L2 order book history (~354 days)
- ~43 GB proprietary Hyperliquid raw market data
- ~26 GB wallet fill history (~90 days)
- ~7 GB HIP-3 market data
- ~330 GB total historical market data, including external sources
Execution Venue: Hyperliquid exclusively
Research Infrastructure: Redis · PostgreSQL/TimescaleDB · QLib · llama.cpp
The objective is not to predict every market movement — it is to identify statistically defensible, economically executable alpha.
- Evidence over intuition: Research hypotheses must survive rigorous out-of-sample testing.
- Net alpha over accuracy: Predictive performance matters only when it translates into executable returns after costs.
- Microstructure over indicators: Focus on market mechanics, liquidity, order flow and price formation.
- Multi-source intelligence: Combine independent information sources and measure their incremental predictive value.
- Robustness over complexity: Prefer reproducible, validated models over overfitted strategies.
- Automation with verification: AI agents accelerate research; mathematical and statistical tests determine validity.
Important
Don't trust, verify. I build for and support decentralized, permissionless and private systems — Bitcoin, self-custody, self-hosted infrastructure and end-to-end encryption.
| 🏆 | Listed in the Proton Security Contributors (Nicolas Bohn, 2026) |
| 🌐 | Bug bounty research on on-chain & exchange infrastructure — API, authentication and protocol integrity |
| Quant & AI | Data & Infrastructure | Exploring |
|---|---|---|
| Python QLib llama.cpp Multi-Agent Systems | PostgreSQL TimescaleDB Redis Docker Linux | Rust |