The Engineering Notebook — page 5 of 38

Notes on building with AI, agents & the modern stack.

Deep dives on AI/ML, RAG systems, agent engineering, and senior-engineer architecture decisions.

Claude Code projects code editor laptop screen — illustration for article on Claude Code Projects: Always‑On Developer Tools

Claude Code Projects: Always‑On Conversation, New Team Risks

Claude Code Projects turns agentic coding into a long-running, persistent workspace. That’s productivity. It’s also a new operational risk surface teams need to govern on day 0.

A smartphone screen displaying the Google Gemini app store page with update options Developer Tools

How to Build a Gemini 3.8 Live Voice Agent [2026 Tutorial]

A practical Gemini 3.8 Live real-time voice agent tutorial: full-duplex streaming, barge-in interruptions, async tool calls, ephemeral tokens, and a latency budget you can actually hit.

A computer screen displaying system information and an Ubuntu logo in a terminal window Developer Tools

How to Measure AI Coding Agent Harness Overhead [2026]

Model wars are a distraction. If you can’t measure prompt scaffolding, repo indexing, context packing, tool calls, and test loops, you’re not benchmarking your coding agent. Here’s a reproducible protocol.

an open laptop computer sitting on top of a table Cybersecurity

How to Implement OWASP Agentic Top 10 Controls [2026]

A control-by-control guide to OWASP agentic top 10 controls: where to gate tool calls (MCP/client/server), what to log, what to block, and what belongs in CI vs runtime.

A computer monitor sitting on top of a desk AI and Machine Learning

How to Run Qwen 35B on 16GB VRAM [2026]: Flags + Quants

A reproducible 16GB recipe for “Qwen 35B”: which Qwen2.5-32B quants fit, how to budget KV cache, and the exact serving flags that stop OOMs.

inference server rackmount gpu machine — illustration for article on How to Serve a Local LLM Cloud and DevOps

How to Serve a Local LLM to Multiple Users [2026]

If your local LLM server “works” but falls apart at 30–50 concurrent chats, this guide shows the real limiter (KV cache) and the exact knobs in vLLM, SGLang, and TGI that move P99.

Computer screen displaying code and project files Developer Tools

Wayback Machine Rate Limiting [2026]: Build a Mini-Wayback

Archive.org’s 429 blocks are real now. Here’s a pragmatic runbook to capture critical URLs, store WARC/WACZ, index snapshots, and replay them in a UI your team will actually use.

A digital dashboard displaying marketing metrics including CTR and quality score on a screen AI and Machine Learning

How to Do Non Deterministic AI System Testing [2026]

A release-gating playbook for LLMs and agents: golden sets, metamorphic tests, variance control, tolerance bands, cohort diffs, and eval-regression postmortems.

black iphone 4 on white table Developer Tools

Single File Web App SQLite [2026]: Migrations + Backups That Work

“One binary + one database file” is a great deployment model until you hit migrations under load and WAL backups. Here’s the safety-first playbook I actually trust.

a computer screen with a picture of a woman on it Developer Tools

Wayback Machine Alternatives [2026]: Developer Archiving Playbook

Wayback Machine alternatives aren’t just other sites to “view old pages.” For developers, archiving is a supply-chain and ops problem: reproducible docs, mirrored release assets, checksums, and restore drills.

a golden padlock sitting on top of a keyboard Technology

SQLite Production API Concurrency: WAL Mode Limits [2026]

SQLite in production is real. WAL mode buys you read/write concurrency and great latency. The trap is write contention, checkpoints, and backups. Here’s the decision guide, tuning defaults, and a runnable load test + restore drill.

closeup photo of computer code screengrab Developer Tools

Topcoat Rust Web Framework Tutorial [2026]: Auth + DB + SSR

A production-minded Topcoat Rust web framework tutorial: project setup, routes + SSR mental model, Postgres + migrations, auth, forms, profiling, and deploy gotchas.