
The model was the easy part
Building camera analytics for retail taught me that the hard part was everything around the model, from what retailers needed to why our GPU sat at its power cap.
Tag archive

Building camera analytics for retail taught me that the hard part was everything around the model, from what retailers needed to why our GPU sat at its power cap.

What happens to a payment that succeeds 200 milliseconds after your timeout fires? Your graph state...
Asked in an LLM inference systems engineer interview. What they are really testing: whether you can...
Most AI products don't need an agent. The first design decision isn't which model or framework to...
Vertex AI alternatives for custom training: one 24-hour 8 x H100 run priced on seven platforms, what you lose by leaving, and how to keep GCP credits.
A practical walkthrough of running DeepSeek‑V3 in ~200 lines of pure JAX, covering why it matters, step‑by‑step usage, and a seasoned engineer’s perspective on its place in production AI stacks.
Use Feast to retrieve historical features for model training and online features for prediction. Then...
A lot of teams get to the same place with their first production model: a cron job retrains it every...
Deploying a machine learning model to production is exhilarating. Your pipeline is live, predictions...
Nvidia is allegedly funding AI startups to buy up its GPUs, creating scarcity. I break down why this matters, practical ways to mitigate the impact, and my perspective as an AI infrastructure engineer.

MLOps: models that survive production · Chapter 4 of 12 · AI Engineering · new chapter every...
Anthropic's IPO brings stable pricing, SLAs, and better tooling for Claude models. I break down what the IPO means, why developers should care, and show practical steps to integrate Claude via API, Terraform, and Kubernetes.