At notYC, I shipped a production AI reviewer for Y Combinator applications, with a retrieval pipeline over 100GB+ of application data. The system earned direct validation from Garry Tan, CEO of Y Combinator.
As a Founding Engineer Intern, I worked end-to-end across data engineering, retrieval architecture, and frontend to bring the product to production. I also built an automated section-by-section scoring engine for application feedback.
In my Security Auditing SLM project, I fine-tuned Llama 3.2 (1B) for private, on-premise security-log auditing and achieved 80%+ factual accuracy on held-out validation logs. I added API-key authentication and rate-limited access through Redis.
My other deployed projects include a multi-model research system that reduced task latency from 8s to 3.2s, and Ledger AI, where a LangGraph workflow reduced query-to-answer time from 6 minutes to 2.1 minutes. I also completed an applied AI/LLM engineering program through Capgemini (EduBridge Program).
