Patrick Pena

I’m a fullstack software engineer focused on designing and shipping reliable, human-centered AI systems. I’ve been building AI-adjacent and decision driven software since 2017, starting with rule based decision-tree systems and data pipelines, and evolving into production LLM and agentic systems as the field matured.

I’ve built and operated scalable cloud backends, applied-AI pipelines, and intelligent agent architectures that support real workflows — from production APIs and observability to stateful, resumable systems and export pipelines.

My work spans TypeScript, Next.js, MCP (Model Context Protocol), Python, FastAPI, PostgreSQL (Supabase/Neon), Redis, Fly.io microVMs, and modern LLM integrations (OpenAI, Anthropic, Perplexity), with a strong emphasis on durability, clarity, and long-term maintainability.

At a high level, I’m interested in AI that supports people over time — systems that are dependable, understandable, and designed to augment human judgment rather than replace it.