Ten years building scalable backend systems and AI-enabled products in Ruby, Python, TypeScript and Go — with a track record of impact in performance, reliability and product delivery.
Linear-style task manager with an AI assistant for conversational queries and autonomous actions — insights, task creation, messaging. Built on an event-driven architecture with a hybrid agentic approach: function calling for structured operations and RAG with pgvector for semantic retrieval.
Conversational appointment booking with an event-driven architecture. An AI assistant powered by LangChain and RAG (pgvector) handles WhatsApp automation and multi-channel support — meeting users wherever they already are.
Built evaluation and testing infrastructure for Fin AI across Rails, TypeScript / React and Python. Shipped core Test page capabilities used by 50%+ of adopting companies, launched Impersonation to open beta for thousands of customers, and improved Messenger reliability for millions of end users by fixing four critical bugs.
Led a critical investment data pipeline with Sidekiq Batches, reducing end-to-end processing time by 90%+. Implemented Datadog monitoring on AWS, designed a real-time funds calculation service that increased processing speed by 50%, and achieved 99.2% test coverage for the Holdings API.
Built and launched a full-stack exam platform with Elixir / Phoenix + LiveView, serving 6,000+ paying doctors in Brazil. Delivered registration and payments, secure exam distribution with end-to-end encryption, and an admin portal for exam operations — then deployed and operated it on AWS with PostgreSQL.
Progressed from Senior Software Engineer to Lead Engineer, leading five developers across four projects. Drove backend delivery with Ruby on Rails, introduced GraphQL across multiple products, and integrated REopt Lite API and NREL SAM to expand renewable-energy capabilities.