Projects

Retrieval-Augmented Generation system built for EECS coursework and research applications.

Studying Pareto frontier between user engagement and lifecycle carbon footprint in large-scale recommendation systems. BPR and SASRec models with carbon-aware re-ranking.

Open-source benchmark harness for comparing LLM inference performance across providers and configurations. Latest vLLM finding (single-worker local setup, Qwen2-0.5B-Instruct): higher concurrency mostly increased queueing latency rather than throughput; doubling max_tokens roughly doubled latency while tokens/sec stayed similar, suggesting generation rate as the bottleneck.

Automated pull request review tool powered by LLMs.

Built an automated LLM-based testing agent to fuzz and validate AI-generated web applications.

Multi-stage LLM agent pipeline to parse freelance job ads, assess feasibility/margins, generate bids, execution plans, and autonomously produce deliverables.

Production AI system to automate building-code compliance checks. Distributed backend with AWS Lambda, API Gateway, S3, Bedrock. Multi-intent routing, caching, auth, guardrails, and monitoring.

MuZero reinforcement learning agent for Snake. Full training pipeline including MCTS, neural network policy/value estimation, and evaluation framework.

ML models to predict maritime vessel trajectories using historical AIS data. Feature engineering pipeline and evaluation across multiple modeling approaches.