A constellation of my work. Follow a star to open it.
- Interpretability & Reasoning
- Agents & Multi-agent Systems
- Computer Vision
- Applied ML & Signals
- Reinforcement Learning
- Robotics
About
AI Engineering student with project experience spanning machine learning, computer vision, reinforcement learning, quantum computing, and LLM agents and interpretability, across healthcare, sports, agriculture, and robotics.
Through academic projects and hands-on experience, I have strengthened my problem-solving, teamwork, and communication skills. I enjoy working at the intersection of technical rigor, experimentation, and product thinking, and I am particularly motivated by opportunities where AI can generate measurable impact.
Stack
Experience
Experience
- Apr 2026 – Aug 2026AI Engineer InternIBMPart-time · Client Engineering (watsonx) · Buenos Aires, Argentina · Hybrid
On IBM's Client Engineering team, built AI solutions with IBM watsonx alongside clients.
- Aug 2025 – Jul 2026Teaching Assistant · Computer Architecture & Operating SystemsUniversidad de San AndrésIn person · Victoria, Buenos Aires
Supported students through low-level systems and C++: architecture, memory, and OS concepts.
- Jan 2026 – Feb 2026Summer Job · Digital Transformation & Continuous ImprovementSyngentaTemporary · In person · Venado Tuerto, Santa Fe
Software and computer-vision work on the digital transformation and continuous-improvement team.
- Aug 2025 – Dec 2025Teaching Assistant · Numerical Methods & OptimizationUniversidad de San AndrésIn person · Victoria, Buenos Aires
Guided students through numerical methods and optimization coursework.
- Aug 2022 – Sep 2022Educational InternshipEssenInternship · In person · Venado Tuerto, Santa Fe
Education
- 2023 – PresentEngineering in Artificial IntelligenceUniversidad de San Andrés
One of the first AI Engineering programs in South America. Coursework across machine learning, deep learning, reinforcement learning, computer vision, robotics, NLP, and systems.
- Sep 2026 – PresentExchange · Coursework in the MSc in Artificial IntelligenceUniversità Bocconi
Honors
Hackathons and competitions.
Quantum key distribution protocols in a realistic eavesdropping scenario
We won the first quantum-computing competition in Argentina. After a month of intensive Qiskit workshops covering fundamentals, circuit design, and cryptographic protocols, we implemented quantum key distribution from first principles. Our project focused on BB84: a complete Qiskit implementation with noise models, error correction, and post-processing reconciliation. We then built a proof-of-concept E91 protocol (entanglement-based QKD using Bell states) to contrast approaches and explore the security properties under eavesdropping attacks, showing how to detect and measure an eavesdropper's presence on the quantum channel.
Winner, AI & Automations category
Delphi, a predictive-intelligence platform for maize-pest outbreak risk, won its category, built on real INTA trap-monitoring data with explainable SHAP-based predictions.
2nd place, stereo vision and edge control for industry
A hackathon organized by Siemens and the Argentine Industrial Union (UIA), tackling real national-industry challenges in energy efficiency, predictive maintenance, cybersecurity, and process optimization for the oilseed sector. We built a stereo-vision, sensing, and edge-control system that measures fragment distribution and thermal conditions to recommend and automate adjustments to the cracker machine's roller opening and VSC conditioning, optimizing energy use and process quality.
Verbalize, interpretability for AI safety
An AI-alignment tool that audits whether a model's stated reasoning matches what it is actually computing, decoding its internal activations into natural language in real time to catch misalignment that chain-of-thought monitoring misses.
SpineAgent, agent over a business's spine object
An autonomous agent that operates on the root operational object crossing an organization's functional areas, with pgvector memory and human-in-the-loop approvals.
CanSat
CanSat is an international student competition where teams design, build, and operate a satellite the size of a soda can. We scored 92.36 at the Preliminary Design Review (PDR) stage, 3rd in Argentina as Team SARF, and we're working on the project again for next year.