Field Notes

Essays from
the frontier.

What actually works, what fails in production, and what I am still figuring out — across AI engineering, autonomous systems, sports analytics, and the path from research to shipping.

Published
What agentic AI actually means in production — beyond the buzzword
AI & Machine LearningMay 19, 202612 min read

What agentic AI actually means in production — beyond the buzzword

Agentic AI is not magic autonomy. In production, it is a controlled loop where a model can plan, use tools, observe results, and decide what to do next inside strict operational boundaries.

AIAgentic AIProduction Systems
Read →
Coming to Field Notes
AI & Machine Learning5 upcoming
How to build a UAV that navigates without GPS — lessons from real research
Zero-shot deployment explained: how meta-learning lets AI generalize without retraining
Why AI systems fail: a taxonomy of 37 real-world breakdowns every developer should know
+2 more
Sports Analytics12 upcoming
How to architect a sports analytics platform from scratch — a practical guide
Why raw stats lie: how context-adjusted metrics make better coaching decisions
How to turn complex sports data into dashboards coaches actually use
+9 more
Future-Focused & Emerging Tech4 upcoming
Quantum computing explained for software engineers — no physics degree required
Quantum machine learning: where quantum computing and AI are starting to intersect
What drones, self-driving cars, and robots have in common — and where they're all heading
+1 more
Career, Research & Learnings5 upcoming
How to go from computer engineering student to AI engineer — a non-linear path
How to write your first undergraduate research paper — a practical step-by-step guide
What a hackathon win actually teaches you that coursework never will
+2 more
Broader Tech & Society4 upcoming
37 times AI went wrong: patterns every developer should recognize before it's too late
Who is responsible when an autonomous system makes a deadly mistake?
How being multilingual makes you a better software engineer than you'd expect
+1 more
Philosophy of AI7 upcoming
Why blindly trusting your AI model is the most dangerous thing you can do as an engineer
The line between AI that helps and AI that decides — and why it matters enormously
Full autonomy isn't always the goal: the case for keeping humans in the loop
+4 more
Building Things in Public5 upcoming
How to build and ship a full-stack sports analytics platform as a solo developer
The solo developer's production stack in 2026 — what actually works and why
I used Claude, GitHub Copilot, and Devin on the same project — here's what I found
+2 more
Tech Career & Student-to-Engineer7 upcoming
What internships actually teach you that no university course ever will
How to land undergraduate research roles — even before you have publications
What transfers from research to industry and what completely falls apart
+4 more