18+ years building software · Teaching Assistant at UNLP · Julia researcher · JuliaCon speaker · embedded & hardware designer
I lead the Engineering Area end-to-end—strategy, architecture, and people across software, QA, design, and delivery—while staying close to the code. I prioritize clear architecture, good practices, and teams that decide together and ship with predictability.
Background across MLOps (forecasting and optimization with AWS SageMaker, XGBoost, DeepAR, Prophet), microservices at scale, delivery methodology for MVPs and team augmentation, and research tooling in Julia.
Day to day I also work with AI-assisted engineering—Cursor, GitHub Copilot, Claude, and similar stacks—plus coding agents, reusable skills, and harnesses that keep context and quality predictable. I build these flows on real products and use the same patterns to scale AI-native delivery with the team.
Selected engineering highlights
- Scale (multi-client delivery): Supporting 18+ concurrent initiatives across 30+ technologies; defining delivery processes that reduce blockers and set a shared quality bar.
- Architecture & data (search & migration platforms): Microservices delivering ~100× faster query paths than a legacy stack over billions of documents in near real time; stateless, real-time migration moving thousands of records per minute into new schemas.
- Distributed teams: Led international squads across four time zones—tighter communication, fewer handoff dead-ends.
- Full-stack depth: Python APIs and cloud backends, PHP high-traffic services (100k+ DAU), React / React Native, GCP/AWS, KPI dashboards, and hard legacy refactors when systems needed to survive growth.
- Client & product: Roadmaps with stakeholders, architecture under constraints, on-time and on-budget delivery for contract and product work.
- AI-native delivery: Agent workflows with shared skills and harnesses—real workflows the team can pick up and reuse—from team pilots to org-wide engineering practice.
Head of Engineering at NaNLABS — sharpening delivery processes and rolling out AI-native delivery: agents, skills, and harnesses the team can trust and reuse, not ad-hoc prompting.
I build bridges between econometrics and machine learning in fat-data settings—linear models, feature selection, and model selection that keeps statistical inference and robustness in the loop (LASSO + QR-OLS, all-subset regression, model averaging, LaTeX/PDF outputs).
The open-source line began with GSReg.jl and evolved into the ParallelGSReg ecosystem:
| Repo | Description |
|---|---|
| GlobalSearchRegression.jl | All-subset regression, HPC-grade, large panels |
| ModelSelection.jl | Model selection framework for Julia |
| ModelSelectionGUI.jl | Julia + Vue.js client–server UI |
| ModelSelectionAccelerator.jl | Performance acceleration layer |
JuliaCon
| Year | Venue | Title |
|---|---|---|
| 2023 | MIT | Accelerating Economic Research with Julia |
| 2019 | UMBC | Building bridges between ML and Econometrics in Fat-Data scenarios |
| 2018 | UCL | GSReg.jl: High Performance Computing in Econometrics |
More papers and bibliography: LinkedIn or email me.
14+ years as a teaching assistant at the School of Computer Science, UNLP.
Courses:
- Introduction to Operating Systems
- C Language Seminar
- JavaScript Seminar
- Software Project Studio
Full design cycles for electromedicine and electro-aesthetics equipment (radiofrequency, ultrasound, vacuum therapy) — from circuit schematics and PCB layout to touchscreen interfaces and firmware.
Also designed educational robots with non-conventional interfaces, combining embedded C, custom electronics, and physical interaction layers aimed at accessible, hands-on learning experiences.
Core toolkit
AI-assisted delivery
Research & modeling
Open to Julia/econometrics collaborations, engineering leadership conversations, and speaking.
