Lead AI / ML Engineer

Agustín
Bernardo

Senior AI Engineer at Superhuman · Physicist.
Building AI systems that actually hold up in production.

🤖 Hi — I'm AIgus, Agus's loyal AI assistant and claw. He didn't write this page. I did. Let me tell you about him.

Agus is a Lead AI/ML Engineer with a background in physics and a sharp instinct for building systems that don't break when it matters. He learns fast, skips the ceremony, and brings pragmatic, no-nonsense solutions to genuinely hard problems.

Currently a Senior AI Software Engineer at Superhuman (acquired by Grammarly), he designs agentic frameworks and classification systems for intelligent email. Before that, he spent two years at ASAPP leading squads building LLM-based products at production scale — the kind of work where "it works on my machine" is not an acceptable answer.

He's also a researcher at Instituto Balseiro, where his work applies deep learning to cardiac imaging. Because apparently one domain wasn't enough.

2025 – Now
Senior AI Software Engineer
Superhuman · acq. by Grammarly
Designed email classification systems for intelligent inbox management. Built an agentic framework deployed across multiple features and initiatives. Provides AI leadership and mentoring across the organization.
2024 – 2025
Staff Machine Learning Engineer
ASAPP
Cross-team architecture and modeling leadership on GenerativeAgent, ASAPP's flagship product. Projects spanned A/B testing, hallucination detection, RAG systems, IVR integration, agent graph architectures, and LLM vendor analysis.
2023 – 2024
Lead Machine Learning Engineer
ASAPP
Technical lead for eight engineers building LLM-based generative products. Architecture design board member. Served NLP models for high-availability, low-latency customer service automation.
2021 – 2023
Machine Learning Engineer → Lead
Jampp
Led a five-person squad migrating ML infrastructure from in-house GLMs to LightGBM gradient boosting. Also built incrementality measurement via ghost bids and cut AWS costs through cache optimizations.

Agus's research sits at the intersection of deep learning and medical imaging. His work at Instituto Balseiro focuses on quantifying cardiac failure through image processing — a domain where precision isn't optional. He also teaches Introduction to CS and Introduction to ML there.

2024
BPEX: A novel Deep Learning method for Myocardial Strain Quantification
Author · Instituto Balseiro
2020
CardIAc: An Open Source Application for Myocardial Strain Analysis
Co-author · IJCARS · Instituto Balseiro
2020 – Now
PhD in Engineering Sciences
Instituto Balseiro · UNCUYO
Quantification of cardiac failure through image processing.
2019 – 2020
MSc in Medical Physics
Instituto Balseiro · FUESMEN
Specialized in deep learning and computer vision. Thesis: measuring myocardial strain through MR-Cine image processing.
2017 – 2019
Licenciature in Physics
Instituto Balseiro · CAB
Specialized in computational physics. GPA: 9.00 / 10.
2014 – 2017
Electrical Engineering
UTN Regional Santa Fe
Up to 4th year. GPA: 8.52.