AI Engineering5 posts

Practical insights and engineering guides on local LLMs, edge computing, secure agent runtimes, and scalable ML infrastructure.

[ COURSE.SYS ]
Cover Image for AI Engineering Course: From Tokens to Canary Rollouts
AI EngineeringCourse · 10 Lessons
transformerslorakv-cachevllmagentsevalswebassembly
Interactive course⋅10 lessons + Python labs

AI Engineering Course: From Tokens to Canary Rollouts

Ten lessons cover tokenization, LoRA, KV caches, MoE routing, PagedAttention, retrieval memory, tool harnesses, judge calibration, tail latency, and canary rollouts. Run the WebAssembly kernel simulator in your browser and the Python exercises locally.

Dan
Dan
[ ARTICLE.SYS ]
Cover Image for Running Codex on a local model with LM Studio
AI Engineering
⋅5 min read

Running Codex on a local model with LM Studio

Most of the jobs my orchestrator hands to Codex are small ones that do not need a frontier model, so I wired the Codex harness to a Qwen MoE running on my own Mac through LM Studio. The wiring was less obvious than expected, and this blogpost collects everything that broke on the way to a working setup, plus the DSPy-style loop it makes affordable.

Dan
Dan
[ ARTICLE.SYS ]
Cover Image for Running time-series forecasting in the browser with Rust and WebAssembly
AI Engineering
⋅5 min read

Running time-series forecasting in the browser with Rust and WebAssembly

For WaySightAI I set the constraint that raw time-series data never leaves the user's browser, which meant the CSV parsing, cleaning, stationarity tests, and forecasting all had to run client-side. The core math is written in Rust and compiled to WebAssembly, and this blogpost goes through the architecture, the benchmark numbers, and the tradeoffs.

Dan
Dan
[ ARTICLE.SYS ]
Cover Image for Orchestrating a forecasting pipeline with Bedrock AgentCore and SageMaker
AI Engineering
⋅5 min read

Orchestrating a forecasting pipeline with Bedrock AgentCore and SageMaker

A forecasting system where an LLM agent runs the whole lifecycle: exploratory analysis in a Code Interpreter sandbox, feature engineering, Bayesian hyperparameter tuning and training on SageMaker, and deployment behind a live endpoint. This blogpost goes through the three-layer architecture and the patterns that keep LLM-generated code contained.

Dan
Dan