Posts

Mapping Motion: Architecting an End-to-End MLOps Pipeline for Legged Robots

1. Introduction: The Sim-to-Real Challenge In traditional Machine Learning, a bad prediction usually results in a lower click-through rate or a flagged email. In robotics, a bad prediction results in physical damage: a robot losing its balance, crashing into a wall or stripping a gear. For an MLOps engineer, this shift in stakes changes everything. The challenge is no longer just about serving a model via an API, it is about managing a lifecycle where the code must interact with the immutable laws of physics.

February 6, 2026 Read

5 Experiments Later: Is Promptfoo the LLM Quality Gate MLOps Has Been Missing?

Quick Introduction At a recent AI meetup in Zurich, a Google engineer put words to a problem I keep seeing in LLM projects: “Not every prompt works well with all providers, and not every tool works well with every provider.” Anyone who’s shipped an LLM application knows this. The issue isn’t awareness; it’s that most teams have no systematic way to evaluate these tradeoffs before they hit production. Just “vibes” and the classic excuse: “We don’t have time to test, business needs it in production this week.”

December 24, 2025 Read

AI Tools Are Changing How We Code, but Who's Teaching Us to Stay Safe?

Artificial intelligence tools are revolutionizing productivity, especially in coding and development. Tools such as GitHub Copilot, Cursor, and Windsurf are reshaping workflows, making software creation faster and more intuitive. However, with these advancements comes significant responsibility, particularly regarding information security. Before you paste that next prompt, do you really know where the data will end up and who might access it? On platforms like YouTube or Medium, countless AI content creators enthusiastically highlight the exciting features of new agentic coding tools. They emphasize their capabilities and the efficiency these tools bring to be more productive. However, this trend often overlooks a crucial issue: the potential risks and consequences of information leaks and compromised security.

June 15, 2025 Read

Why I Replaced Poetry with uv for almost everything

I used to be that person who would smugly paste a poetry install command into every Dockerfile, CI, etc. I even wrote a blog post last year praising Poetry—it still felt like the obvious default—until one night, after skimming a few posts about uv and waiting on slow builds after every tiny tweak, I installed it out of curiosity. The build that had been taking two minutes finished in fourteen seconds. 🤯

May 31, 2025 Read

Unlocking Business Insights With LLMs

Three practical architectures for exploiting Large Language Models With the “boom” of artificial intelligence, specifically with in the field of natural language, more people and companies are interested on use cases to improve their business with them. In this article I tried to give you some architectures, one more easy to implement than others, to help you understand and inspire to build your own solutions. 1. Using embeddings for semantic search Using a mathematical representation of words and semantic search, you can build a solution using an Large Language Model (LLM) in just a few steps to improve the relevance and precision of interaction with your own data.

February 18, 2024 Read