About Me

• Built and hardened production integrations with three external systems (Bullhorn ATS, Gmail, Microsoft Graph) behind a swappable provider interface — reverse-engineering undocumented behavior (a silent sender-format quirk, an unstated OAuth scope requirement) that no public docs covered.
• Built a hybrid retrieval and prompt-injection-hardened LLM pipeline that cut inference costs ~80% while screening and routing 1,000+ applicants per posting via LLM scoring and reply-intent triage.
• Designed a Postgres-backed reliability layer so multi-step agent workflows stayed correct even when individual steps failed.
• Led a 4-person engineering team shipping a 3-service, ~20K-line production platform (Next.js frontend, FastAPI backend, PostgreSQL, 310+ unit tests); partnered directly with enterprise customers through weekly demos and technical reviews.

• Built the initial prototype of an autonomous recruiting agent (LangGraph, MCP) integrating the Bullhorn ATS, Gmail, and Google Calendar APIs to automate resume ingestion through interview scheduling.
• Established the agent architecture, model-routing design, and service boundaries that became the foundation of Gentoro's production recruiting platform.

• Built a data visualization platform that cut optical lens comparison time from 2+ hours to 5 minutes; presented to engineering leadership, driving adoption across multiple teams.
• Developed Python automation tools to process inspection data from semiconductor manufacturing systems, reducing manual analysis time and improving workflow efficiency.
• Collaborated with hardware and software engineers to identify bottlenecks and design scalable internal tooling.

Relevant Coursework: Large Language Models, Software Engineering, Automata Theory, AI, Machine Learning.

Did a study abroad to get a chance to learn from a top 25 CS Institution in the world.

Supplemental coursework alongside Occidental College studies.