About
I'm an AI/ML engineering leader, AI-native builder, and researcher. A decade leading and shipping ML systems across a federal transportation regulator, a PhD lab, Google, and Amazon. I'm now leading AI & ML at Wizard AI and founding Athlete Space on the side.
Position
Three identities, one practice.
Today I'm Leading AI & ML / Principal Scientist at Wizard AI, where I lead a 5-person engineering and ML team delivering an end-to-end agentic AI platform. I own the technical roadmap, model and orchestration choices, the evaluation stack, the responsible-AI posture, hiring, and the technical narrative we take to customers and investors.
I'm also founder & CEO of Athlete Space, an AI-native endurance-training platform. Athlete Space is where my research background and my operator instinct meet: a production multi-agent coaching system grounded in objective training-load science, with real-time pose-estimation form analysis, built solo from architecture to ship.
Before that I was an Applied Scientist II at Amazon, an AI & ML Researcher at Google, and a PhD candidate at RIT publishing on human pose estimation and multi-scale vision architectures. Before everything, I was a software engineer in applied ML at Transport Canada, which is where I learned that models only matter when they clear a real decision.
How I lead
Rigour is velocity.
I was trained as a researcher before I was trained as a product person, and it shows. I push for evaluation harnesses, typed contracts, feature flags, and staged rollouts. Not because it's slower. Because it's the only way teams ship AI at pace without quietly eroding trust.
Production is the laboratory.
The hardest problems in AI right now aren't in the model. They're in orchestration, cost, latency, eval, and trust. I've spent years on the engineering side of the ML boundary, and I hire and build teams that can hold both sides.
Own the narrative.
A technical leader's job is not just the team and the stack. It's the story. Boards, customers, and regulators don't need jargon; they need a crisp technical thesis, a risk model, and a cadence they can plan around. Every role I've had, I've been the translation layer between AI capability and business decisions.
Respect the domain.
Strong applied AI work is built on strong domain grounding. For Athlete Space, that's exercise physiology. For industrial ML, that's ops. I don't try to win arguments with domain experts. I codify their knowledge into the system and the eval.
Education
2018 – 2022
Ph.D.
Electrical & Computer Engineering
Rochester Institute of Technology
GPA 3.96 · Advised by Prof. Andreas Savakis (Computer Vision Laboratory). Thesis: Multi-Scale Architectures for Human Pose Estimation.
2016 – 2017
M.Eng.
Electrical Engineering
Memorial University of Newfoundland
GPA 4.00 · Thesis: UAS Integration to Airspace and Collision Risk Assessment.
2010 – 2015
B.Eng.
Electrical Engineering
UNESP, São Paulo State University (Brazil)
Exchange year at University of Toronto.
Outside the work
Endurance is the through-line.
Sub-2:30 marathon. Ironman champion. A decade of structured endurance training every week, through PhD and Big Tech and into venture. Athlete Space exists because I am the user.
I live in Nashville with my wife and our two daughters.
