Where the mathematics mattered into building something!

A career that runs from stochastic-optimization research into building real systems, and back!

  1. LabMA — UFRJ

    2025 — Present · prev. Research Fellow 2023–24

    Consultant

    Lead data-centric projects on a 1TB+ Oracle base covering ~130 million Brazilians (1B+ records). Defined scope, managed an analyst team, and drove a multi-engine benchmark (PostgreSQL, Hydra, ClickHouse, DuckDB) for the biometric-tables (BR-EMS) pipeline — execution-plan analysis, bitmask techniques, multi-year ingestion on pgduckdb. Introduced versioning, modularity and QA discipline lab-wide. I first arrived here as an undergraduate research fellow.

    // my takeaway from the role

    LabMA is where i learned to manage a bunch of people, a learned a lot about that stuff. It's fun learning to design first how to handle 50tb of data, how to optimize for a insane dataflow, and then learn that you... kinda don't need that if you don't need that. Optimize for what you have, and learn how to make stuff FAST and GOOD.

    • Oracle
    • DuckDB
    • pipelines
    • leadership

  2. Pluma Finance

    Feb 2026 — Apr 2026

    Founding Engineer

    Founding engineer at a personal-finance fintech. Designed and built an autonomous multi-agent pipeline that takes software work from ticket to merged PR — specialized agents (implementer, reviewer, fixer, custodian) cooperating across iterative rounds, with multi-model cross-review (Claude + Gemini) and containerized guardrails. Also built the transaction pipeline (webhooks, dedup, recurring-transaction detection, credit-card engine) and tuned hot PostgreSQL query paths. Wore the Product Owner hat for part of the run.

    // my takeaway from the role

    I had a lot of fun at Pluma! definitely a cool place to work in, and i feel that i developed a lot in my work skills there. Had to leave because of time constraints: a PHD takes more time than i thought :x

    • TypeScript
    • PostgreSQL
    • tRPC
    • AI agents

  3. NACAD — UFRJ

    2024 — 2025

    Research Assistant

    Built a diffusion-based super-resolution model that sharpens reduced-order fluid-dynamics output (published at CILAMCE 2024). Refactored the reduced-order model from a Jupyter notebook into a documented Python module with safe configuration loading.

    // my takeaway from the role

    NACAD was where i played a lot with computational fluid dynamics, machine learning models and learned to refactor gigantic notebooks into modules.

    • PyTorch
    • diffusion
    • CFD

  4. Bee Datascience

    2024

    Software Developer · Contractor

    Built and deployed a portfolio-generation system based on non-linear optimization of a target metric for market trading, and integrated the strategy into an automated trading system. Further details under NDA.

    // my takeaway from the role

    Bee datascience was my first professional experience. I worked building a portfolio-generation system, which was very fun! i also learned to organize my work.

    • Python
    • optimization
    • trading

  5. Institute of Mathematics — UFRJ

    2022 — 2023

    Undergraduate Research · Stochastic Optimization

    My first research chapter, in partnership with the national system operator (ONS): implemented stochastic-optimization methods in Julia (SDDP.jl, JuMP.jl) for the Brazilian hydrothermal dispatch problem, with reproducible Docker builds. The seed of the current PhD.

    // my takeaway from the role

    My first research project, and my first taste of academia. To say that this project changed my life is to say very little about it. It is, and was, the first step on a great journey, and i still look back and think about some of the lessons i learned from there.

    • Julia
    • SDDP
    • ONS