Carlos E. TorresBlog
I use this space to document what I’m building, analyzing, learning, and studying. I don't have a strict schedule for posting, but I do have a running list of topics I'm excited to write about. From web and mobile development, to clean systems architectures, to AI, ML, LLMs, and intelligent, human-centered systems. Hope you find something useful here.
Feel free to 
Subscribe.
Accuracy Is the Wrong Metric for Clinical AI Agents
August 11, 2026 · 9 min readBenchmark accuracy tells you a model can answer questions. It tells you almost nothing about whether a clinician should trust it at 3 a.m. Here is what actually breaks when clinical AI agents move from the leaderboard to the ward, and the five layers I think an evaluation stack needs instead.
Is Fine-Tuning AI Models Still Needed?
August 3, 2026 · 6 min readFine-tuning is no longer the default first move, yet the frontier models we prompt so casually only exist because of it. A practical look at when to prompt, when to retrieve, and when to train.
What is AI Observability?
July 31, 2026 · 3 min readTraditional monitoring tells you if a system is down. AI Observability reveals why your model is misbehaving in the real world. Here is your quick guide to the key pillars, top tools, and trends keeping modern AI reliable.
Human-Centered AI: Putting People at the Heart of Intelligent Systems
July 31, 2026 · 5 min readHuman-Centered AI: Why building AI that amplifies people, rather than replaces them, is the research question at the heart of my PhD journey.