🎓 CS & Applied Math @ NYU • Data Science · AI/ML
B2B AI platform that delivers deal analysis and consulting insights to SMBs.
- Built multi-agent orchestration system with RAG pipelines for document analysis
- Designed vector search layer (pgvector + Supabase) handling structured and unstructured data
- Optimized inference costs 40% via model routing (GPT-4o for complex queries, DeepSeek for routine tasks)
Tech: FastAPI, PostgreSQL, Supabase, pgvector, LangChain
Predicting loan default with production-realistic constraints.
- Trained and evaluated models on 1.3M+ LendingClub loans with strict leakage prevention
- Addressed class imbalance and benchmarked logistic regression, random forest, and gradient boosting
- Translated probabilities into decision thresholds (approve / review / reject) rather than raw scores
Tech: Python, pandas, NumPy, scikit-learn
Business question → structured analysis (JSON, spreadsheet, deck).
- Converts vague prompts into explicit analysis plans using agent graphs
- Retrieval-grounded generation to ensure outputs are traceable to source data
- Benchmarked cost and latency tradeoffs across routing strategies
Tech: LangGraph, FastAPI, Redis, Docker
Framework to distinguish real trading signal from noise.
- Deterministic backtests with transaction cost modeling
- Statistical validation via bootstrap CI, permutation tests, and Monte Carlo simulation
- Explicit verdicts: edge, inconclusive, or noise — not just Sharpe ratios
Tech: Python, pandas, NumPy, SciPy, Streamlit, pytest
NLP + hypothesis testing on lyrical evolution.
- Sentiment analysis, lexical diversity metrics, topic modeling
- Statistical tests across albums (not just visualization or anecdotes)
Tech: Python, scikit-learn, React, Vite
Languages: Python, SQL, Java, C++
Data: pandas, NumPy, SciPy, scikit-learn, dbt
Infrastructure: AWS, FastAPI, PostgreSQL, Redis, Docker, Prefect
ML/AI: LangChain, LangGraph, RAG pipelines, vector databases
💡 I like building things that actually work, then figuring out why they work.