> cat README.md

last updated Jul 2026

# Hi, I'm Vicente

I design and build agentic systems — software that reasons, plans, and acts on its own. Part engineer, part tinkerer, wholly obsessed with making machines genuinely useful.

explore.agents()$ say-hello
- role
AI Engineer
- focus
Agents / LLMs
- based
Remote
- status
Building

## about

I've spent the last few years building at the edge of what language models can do — turning raw model calls into reliable, tool-using agents that do real work instead of just talking about it.

My approach is simple: small, composable pieces; tight feedback loops; and a healthy skepticism of anything that only works in a demo. I care about the boring parts — evals, guardrails, observability — because that's what makes agents trustworthy.

> “The best agent is the one you forget is running.”

### stack

  • TypeScript
  • Python
  • Next.js
  • AI SDK
  • tRPC
  • Postgres
  • Vercel
  • Rust

## agents

The primitives I keep reaching for. Each one is a building block for autonomous, dependable systems.

  • - reasoning-core

    A planning agent that decomposes fuzzy goals into verifiable steps and self-corrects along the way.

  • - tool-router

    Dynamic tool selection with typed schemas, retries, and graceful fallbacks when an API misbehaves.

  • - durable-runs

    Long-running, resumable agent workflows that survive restarts and pick up exactly where they left off.

  • - eval-harness

    Continuous evals and traces so every prompt change is measured, not guessed.

## work

A few things I've shipped and open-sourced. More on .