5 articles
From dependency hell to device confusion - I built a modality-aware cookiecutter template that gets you from idea to training in minutes, not hours. Here's why modern ML projects need modern tooling.
Master efficient ML development with the Pre-Flight, In-Flight, Post-Flight framework. Learn how to catch bugs before expensive training runs, monitor training in real-time, and evaluate models beyond vanity metrics. Stop wasting millions on failed ML experiments.
A comprehensive overview of modern ML training infrastructure, covering cloud agnosticism, spot instances, on-premise solutions, heterogeneous hardware, distributed training, and emerging GPU cloud providers.
The history of Full stack, Machine learning which results in the mergence of the ML full stack engineer
Problems in the ML ecosystem. Fragmentation in machine learning, that keeps preventing the stack from growing higher. How I took a stab at the issue.