About me

I spent 10 years shipping software the traditional way. Now at numa, I'm changing how we build it, with AI agents involved in every step from ticket to production.

Most of my work is building what makes the agents reliable. That includes test and eval pipelines, LLM deployments on AWS, and workflows that take the repetitive work off engineers' plates. I still own the architecture of our pricing platform, so I see every day where AI helps in production and where it doesn't.

Before numa, I worked in large corporations, consulted for major banks and fintechs, and built products at startups and scale-ups of every stage.

If you're working on agentic engineering or bringing AI into a tech team, feel free to reach out. I also enjoy mentoring.

What I'm doing

  • Agentic SDLC

    A process where agents take work through plan, spec, split and build to a reviewed PR. Bug reports go from triage to an open PR on their own.

  • Pricing Platforms & Algorithms

    Architecture of numa's pricing platform, the algorithm-driven system that sets room prices across our properties every day.

  • Full-Stack Development

    TypeScript end to end: NestJS on serverless AWS for the backend, React for the tools people use.

  • Technical Leadership

    Building and mentoring teams, and helping engineers bring AI tools into their daily work.

Resume

Download as PDF

Education

  1. AI Hero by Matt Pocock

    May 2026 — Aug 2026

    Course covering a broad spectrum of AI engineering patterns — agents, evals, RAG and tool use — in TypeScript.

  2. University of São Paulo — BSc Computer Science

    Jan 2012 — Dec 2017

    Held a CNPq Scientific Initiation scholarship for one year, leading to the publication of the research "Event Simulator for Spacial Systems".

Experience

  1. Agentic Lead — numa (formerly COSI)

    Nov 2025 — Present

    Leading numa's move to AI-assisted engineering. I designed our agentic SDLC, the process every team now uses to ship a change, with agents doing most of the work at each step and engineers making the decisions.

    • Plan, spec, split, build: agents take a problem through a spec and tickets to a green PR, checked against our Definition of Ready
    • Three review layers before merge: an agent code review, automated review in CI, then a human gate focused on what machines can't judge
    • Our own AI model, our agent in Linear, takes a bug report from triage to an open PR on its own
  2. Technical Lead — numa (formerly COSI)

    Mar 2021 — Oct 2025

    Led the engineering team behind numa's pricing platform, the system that sets room prices across our properties every day. We designed the architecture, moved it to serverless AWS, and turned the pricing logic from manual rules into algorithm-driven tools that the revenue team could trust and adjust on their own.

    • Built the platform on serverless AWS with NestJS and TypeScript
    • Built internal React tools for pricing and property go-lives
    • Mentored the team and set the technical standards we still use
  3. Consultant — Devmania

    Mar 2020 — Feb 2021

    A small software factory made up only of highly specialized people, covering engineering, architecture, design and product. Companies like Nubank, Itaú and TON brought us in when they had a product to launch on a tight deadline, and we took it from idea to production fast.

    • Built and shipped web apps and backend services for banks and fintechs
    • Set up the architecture and infrastructure for new products (Node.js, React, Docker, MongoDB)
    • Worked remotely alongside client teams, often under launch pressure
  4. Tech Lead — Learn To Fly

    Feb 2018 — Feb 2021

    First technical hire, brought in to build a team and turn the idea into a working product. I designed the architecture, built most of the system from scratch, and led the developers and designers I hired from the first version to production.

    • Designed and built the platform with React, Node.js and Kubernetes on AWS
    • Hired and led the engineering and design team
    • Set up the engineering basics early: CI/CD, production monitoring and bug tracking
  5. Full-Stack Engineer — Colmeia

    Apr 2017 — Feb 2018

    Full-stack engineer at a startup building a social network as a service, a platform that let companies run their own branded social networks. I led a small team of developers and designers that delivered the whole client app.

    • Built the client app in Angular, designed to scale as the user base grew
    • Helped design and build the Node.js backend on Google Cloud
    • Led the day-to-day work of the developers and designers on the client side
  6. Founder — EduCaqui

    Mar 2016 — Feb 2017

    Founded with friends as a peer-to-peer tutoring platform. Took it from ideation to implementation and raised angel investment, but the company folded during the test phase.

  7. Intern — Itaú BBA

    Aug 2014 — Aug 2016

    Internship on the technology team at Itaú BBA, the investment banking arm of Itaú, working on systems for the bank's international units in New York, London and Nassau/Cayman.

    • Worked on software development and business analysis for international banking systems
    • Gathered requirements with business teams, built screen prototypes and wrote specifications
    • Supported business teams during testing and tracked incidents to resolution
  8. Undergraduate Student Researcher — IME-USP

    Jul 2013 — Jul 2014

    Research at the Institute of Mathematics and Statistics of the University of São Paulo that led to the publication of the paper "Event Simulator for Spacial Systems".

Skills

  • AI
    • AI Agents
    • Claude Code
    • Large Language Models (LLM)
    • LLMOps
    • AI Evaluation
    • Prompt Engineering
  • Stack
    • TypeScript
    • NestJS
    • React.js
    • Node.js
    • Serverless AWS
  • Leadership
    • Technical Leadership
    • Engineering Management
    • Mentoring

Blog

Contact

Contact Form