OpenTeams Engineering Blog https://openteams.com/engineering-blog/ Benchmarks, systems deep dives, and hard-won lessons from OpenTeams engineers. Reliable Visual Regression Testing for Humans and Coding Agents https://openteams.com/engineering-blog/visual-regression-testing-jupyterlab/ https://openteams.com/engineering-blog/visual-regression-testing-jupyterlab/ Tue, 06 Oct 2026 14:00:12 GMT See how JupyterLab made visual regression testing reproducible on Linux machines, cut CI from 55 to 15 minutes and let contributors update snapshots. Your AI Agent Ignored AGENTS.md. Your Linter Won't Let It. https://openteams.com/engineering-blog/lint-rules-for-ai-agents/ https://openteams.com/engineering-blog/lint-rules-for-ai-agents/ Thu, 01 Oct 2026 07:26:51 GMT Use lint rules for AI agents to enforce project conventions. Custom ESLint rules like @jupyter/eslint-plugin give agents clear errors they can fix on their own. Where Does Jev Fit in a Software Supply Chain? We Benchmarked It Against the Open Alternatives https://openteams.com/engineering-blog/jev-bench-package-curation/ https://openteams.com/engineering-blog/jev-bench-package-curation/ Tue, 29 Sep 2026 08:59:43 GMT We benchmarked TypeSafe's Jev, Laya, CLM-8B and Claude Haiku on five package-curation triage tasks. A fine-tuned 421M model matched Jev at a tenth of the latency. Awesome Jupyter AI: A Map of 100+ Jupyter Extensions for AI https://openteams.com/engineering-blog/awesome-jupyter-ai-extensions/ https://openteams.com/engineering-blog/awesome-jupyter-ai-extensions/ Fri, 25 Sep 2026 12:52:56 GMT Find the right Jupyter AI extension: a curated map of 100+ chat panels, inline completers, agent bridges and MCP servers for JupyterLab and Notebook 7. MiniCPM5-2B quantization report https://openteams.com/engineering-blog/minicpm5-2b-quantization-report/ https://openteams.com/engineering-blog/minicpm5-2b-quantization-report/ Wed, 23 Sep 2026 01:18:00 GMT MiniCPM5-2B quantization report: the best GGUF weights and K/V cache quants on llama.cpp and BeeLlama.cpp, squeezing a SOTA model into 3 GiB RAM. I Asked LLMs to Review Another LLM. They Still Got It Wrong https://openteams.com/engineering-blog/llm-review-reliability/ https://openteams.com/engineering-blog/llm-review-reliability/ Fri, 18 Sep 2026 02:24:06 GMT I tested LLM reviewer models on meeting summaries to see when they catch hallucinations, when they delete true claims, and how to choose one. LLMs: Intelligence vs. cost https://openteams.com/engineering-blog/intelligence-vs-cost/ https://openteams.com/engineering-blog/intelligence-vs-cost/ Wed, 02 Sep 2026 01:29:09 GMT LLM intelligence vs cost: logarithmic scales in plots hide the cost chasm between frontier and cheap models; actual OpenRouter prices alter the Pareto frontier. pixi on UBI-micro: A Safer, Smaller Multi-Stage Container Build https://openteams.com/engineering-blog/pixi-ubi-micro-containers/ https://openteams.com/engineering-blog/pixi-ubi-micro-containers/ Fri, 24 Jul 2026 12:21:21 GMT How a UBI-micro multi-stage build shrinks a pixi container's attack surface and CVE count, with side-by-side Trivy and Grype scans on UBI full, minimal, and micro. The Best Code Review Says Less https://openteams.com/engineering-blog/best-code-reviews-say-less/ https://openteams.com/engineering-blog/best-code-reviews-say-less/ Fri, 24 Jul 2026 12:06:01 GMT Far too many AI code review comments are noise. Pair a human editor with AI, surface only the findings that matter, and encode that discipline in a reusable skill. You Don't Drive AI. You Ride It. https://openteams.com/engineering-blog/you-dont-drive-ai-you-ride-it/ https://openteams.com/engineering-blog/you-dont-drive-ai-you-ride-it/ Thu, 09 Jul 2026 21:37:12 GMT AI is stochastic, not deterministic. You don't drive it, you ride it: make documents the memory, use the variation to prototype, and verify every public claim. Who Is Your Code's Audience? An Engineering Team Talks AI-Written Code https://openteams.com/engineering-blog/who-is-your-codes-audience/ https://openteams.com/engineering-blog/who-is-your-codes-audience/ Thu, 09 Jul 2026 21:36:59 GMT Learn why your code's audience decides how much readability matters, as an engineering team debates AI-written code, review, and accountability. The tp_as_number Slot and Binary Operation Dispatch in CPython https://openteams.com/engineering-blog/tp-as-number-slot-and-binary-dispatch/ https://openteams.com/engineering-blog/tp-as-number-slot-and-binary-dispatch/ Mon, 22 Jun 2026 14:01:40 GMT How CPython dispatches binary operations via tp_as_number slots, and how TorchDynamo mirrors that model to improve correctness and reduce ad-hoc special cases. PDF Table Extraction: Docling vs Marker vs LlamaParse Compared https://openteams.com/engineering-blog/docling-vs-marker-vs-llamaparse/ https://openteams.com/engineering-blog/docling-vs-marker-vs-llamaparse/ Wed, 20 May 2026 03:52:48 GMT Compare three Python tools for PDF table extraction: Docling, Marker, and LlamaParse. Learn which handles merged cells and multi-level headers best. Sandboxing Code Mode for Local LLM Agents https://openteams.com/engineering-blog/code-mode-sandboxing-local-llms/ https://openteams.com/engineering-blog/code-mode-sandboxing-local-llms/ Tue, 05 May 2026 03:01:31 GMT Code mode can make local LLM agents more practical, but executing model-written code brings sandboxing back into the architecture. Plugin Playground AI Integration for Faster Plugin Prototyping https://openteams.com/engineering-blog/plugin-playground-ai-integration/ https://openteams.com/engineering-blog/plugin-playground-ai-integration/ Sat, 25 Apr 2026 02:23:05 GMT Learn how we integrated AI into Plugin Playground to help you create, edit, test, package, and share JupyterLab plugins faster in JupyterLite and Binder. Slow Down — Simple Lessons for Guiding AI and Shipping Better Code https://openteams.com/engineering-blog/slow-down-ship-better-code/ https://openteams.com/engineering-blog/slow-down-ship-better-code/ Thu, 23 Apr 2026 08:32:20 GMT Practical lessons for shipping better code, staying in control, keeping your skills sharp, and getting real value from AI coding tools without losing yourself in the hype. We Benchmarked 6 Python Package Managers on a Real ML Project. Here's What We Found. https://openteams.com/engineering-blog/benchmark-python-package-managers/ https://openteams.com/engineering-blog/benchmark-python-package-managers/ Wed, 22 Apr 2026 13:07:27 GMT Head-to-head benchmark of pixi, uv, conda, mamba, pip, and poetry on a real ML/AI project with 25+ mixed conda-forge and PyPI dependencies. What I Learned Making a Local LLM Do Real Work https://openteams.com/engineering-blog/what-i-learned-making-a-local-llm-do-real-work/ https://openteams.com/engineering-blog/what-i-learned-making-a-local-llm-do-real-work/ Wed, 22 Apr 2026 04:05:08 GMT Learn what makes a local LLM agent reliable enough for real work: evals, deterministic Python logic, and why a better model beats engineering. From Skill to Agent: When a Text File Isn't Enough https://openteams.com/engineering-blog/from-skill-to-agent/ https://openteams.com/engineering-blog/from-skill-to-agent/ Wed, 22 Apr 2026 04:05:06 GMT When does a Claude Code skill stop being enough? See why credential security pushes real workflows toward proper agent architectures.