At Trinetix, I build prompt flows and production LLM systems for a Big Four client serving 100k+ monthly active users.
I created and maintained 20+ prompts, improved a critical RAG solution from 20% to 90% accuracy, and reduced sensitive-data report creation from two weeks to two minutes.
I work closely with backend, frontend, QA, business analysts, clients, and C-level stakeholders to ship prompt flows, structured outputs, evaluation datasets, and AI marketing workflows. I also established the team’s first prompt development, evaluation, and documentation standards.
In my Slack AI agent project, I built a provider-agnostic tool-use agent for Anthropic and OpenAI, with RAG on Postgres and pgvector, Redis-backed approval flows, encrypted credentials, and production services built with FastAPI, Celery, and continuous deployment.
