Building production ML/AI systems. MS Computer Science @ Stevens Institute of Technology (May 2026).
I work on retrieval pipelines, agent architectures, and applied interpretability. Currently Data Analyst at Giraffe Media Group, where I build SMS analytics infrastructure on Redshift and ship internal AI tooling. Mechanistic interpretability research on GPT-2 advised by Matthew Finch (Senior ML Engineer, Bloomberg LP).
Open to ML/AI Engineer roles starting May 2026.
Earnings Intelligence Platform
RAG system over SEC 10-K filings with rigorous evaluation across 12 retrieval configurations. Best composite score 0.870 using semantic chunking with hybrid retrieval and a cross-encoder reranker, evaluated via RAGAS on AAPL, MSFT, GOOGL, NVDA, META.
ChromaDB BM25 GPT-4o-mini MLflow sentence-transformers SEC EDGAR API
Self-Healing RL Pipeline
Four-agent A2A architecture (Monitor, Diagnostics, Repair, Verification) using DQN for autonomous recovery from data drift, tested across simulated subreddit phase shifts.
PyTorch FastAPI LangSmith
Job Hunter
Automated job discovery scraping 55+ sources with H-1B sponsorship filtering, match scoring, Google Sheets output, Gmail digests, and push notifications. Deployed on GitHub Actions.
Python GitHub Actions ntfy.sh Google Workspace API
Mechanistic interpretability of decoder-only transformers. Greedy sequential SVD pruning of GPT-2 attention heads achieving 6.15% loss improvement with zero retraining; scaling experiments to GPT-2 Medium.
Python PyTorch LangChain LangGraph ChromaDB FastAPI TypeScript SQL (Redshift, PostgreSQL) AWS Docker GitHub Actions