Digest #46: February 2026
February at Mad Devs is all about local AI, smarter logs, and visual-first learning. This month's digest brings practical experiments with coding agents and local models, plus a new Visual Glossary format that turns core complex concepts into shareable slide decks.
Tech Journal: local AI, real pipelines
This month's Tech Journal focuses on how to use AI in production without turning everything into a cloud billing exercise.
Our tech guide shows how we automated glossary interlinking using FastEmbed and a 130 MB local model on CPU. No API keys, no rate limits, and semantic links that scale to thousands of terms with zero external dependencies.
A practical guide to using OpenCode as a local agent inside your application. You will see how to run the agent as a subprocess, delegate web search and formatting to it, and avoid token juggling, key rotation, and vendor lock-in.
An experiment where production Docker logs become input for Google Jules. The agent analyzes errors and patterns and turns them into feature ideas and design drafts, so noisy logs start generating structured product improvements.
Visual Glossary: concepts in fun slides
This month, we introduce a new category, the Visual Glossary. Each term comes as a short slide deck you can show to your team: real problems, context, and concrete takeaways, not textbook definitions.
A visual overview of the shift from passive models to agents that can plan, act, and react to feedback. The deck focuses on control, accountability, and how to keep autonomy useful in production settings.
Slides that explain when compact models are a better choice than huge ones. The focus is on cost, latency, on-prem and edge deployment, and targeted use cases where "small and sharp" beats "large and general."
A visual guide to how text is turned into vectors and why this matters for search, RAG pipelines, recommendations, and agent reasoning.
Slides that break down what it really means for software to scale: handling more users and data without falling apart or becoming too expensive to run.
Careers: build with AI, not around it
We are hiring across AI, engineering, and operations for teams that work with agents, local models, and production-grade systems.
AI & ML POSITIONS
π AI Agent Developer
π Prompt Engineer
π Junior+ ML Engineer
- - - - -Β - - - - - - - - - - - - - - - - - - - - - -
ENGINEERING POSITIONS
π Senior Frontend Developer (React.js)
π Solutions Integration Engineer
π Middle QA Automation Engineer
- - - - -Β - - - - - - - - - - - - - - - - - - - - - -
OPERATIONS
π Office Manager