M.S. Mathematical Finance & Financial Technology @ Boston University (Questrom) · Graduating Jan 2027
Working at the intersection of quantitative research, deep learning, and LLM post-training — especially where reinforcement learning meets language models.
🌱 Focus: Quant Research · Deep Learning (PyTorch) · LLM × Finance · Reinforcement Learning
🔭 Currently
- Offline RL for portfolio allocation — research project with MathWorks
- Building deep learning foundations from scratch (Karpathy Zero-to-Hero → Transformers → LLM post-training)
- Studying RLHF / DPO / GRPO toward LLM alignment
| Project | What it is |
|---|---|
| picograd | Reverse-mode autograd from scratch — foundations for understanding LLM training |
| mathworks-offline-rl | Offline RL (CQL/IQL/DDPG) for portfolio allocation · MathWorks research |
| llm-event-study-alpha | GPT-4o + RAG on earnings calls → long-short alpha (+1.66%, p=0.029) |
| Bitcoin-Predictive-Signal-Modeling | Minute-level BTC forecasting · CNN-LSTM-Attention · Sharpe 0.33 |
| VIX-Mean-Reversion-Algo-Trading-IBKR | Fully automated VIX futures system on the IBKR API |
Quant: factor modeling · IC/ICIR · backtesting · portfolio optimization ML/RL: deep learning · CNN/LSTM/Transformer · offline RL · XGBoost/LightGBM LLM: GPT-4o · RAG · NLP · post-training (RLHF/DPO/GRPO)
