OpenTeams

The promise of AI coding tools is speed.

But faster does not always mean better, especially when teams start skipping the checks that make code reliable.

This article by Johnny Bouder shows how developers can use AI coding tools to move faster without giving up control, quality, or engineering judgment.

🚀 Link: https://na2.hubs.ly/H05NngT0

Slow Down — Simple Lessons for Guiding AI and Shipping Better Code 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.

Happening today at 1 pm ET.

Proving Model Provenance: EO 14365, NIST AI RMF, and Federal Audit Readiness.

Our own Chuck McAndrew + Carahsoft. Part of our ongoing federal series. CPE eligible.

Register and join → https://na2.hubs.ly/H05NpbJ0

A Claude Code skill can turn a simple markdown file into something that feels like working software.

Adam Lewis explored this with a Harvest time-tracking workflow. The skill described the CLI, onboarding flow, billing preferences, and common commands well enough for Claude Code to log, edit, and delete time entries.

That is powerful because there is no packaging, compilation, or deployment step.

But it also exposes the limit of instruction-only systems. A skill can tell an agent what to do, but it cannot enforce what the agent is allowed to access.

This article explains when a skill is enough, when it starts to fail as a security boundary, and how teams can reduce credential exposure with stricter controls.

🚀 Link: https://na2.hubs.ly/H05KVtG0

From Skill to Agent: When a Text File Isn't Enough When does a Claude Code skill stop being enough? See why credential security pushes real workflows toward proper agent architectures.

Most federal AI environments mix proprietary systems, open-weight models, and open-source tooling.

Proving provenance across all of that, and being audit-ready when the time comes, is hard.

Chuck McAndrew covers this in Session 3 of our series with Carahsoft. May 28th, 1 pm ET.

Register today → https://na2.hubs.ly/H05KX__0

If you ask which Python package manager is best, you will usually get a strong opinion.

But the better question is: best for what?

Brandon Geraci compared `uv`, `pixi`, `conda`, `mamba`, `pip`, and `poetry` on a realistic ML project with 25+ direct dependencies and 200+ transitive packages.

They were evaluated across:

• Install speed
• Lockfile generation
• Conda-forge compatibility
• Disk footprint
• Mixed conda/PyPI dependency behavior

The interesting part is that the fastest tool is not always the best fit.

This article breaks down the trade-offs and explains when each package manager is the right fit.

We Benchmarked 6 Python Package Managers on a Real ML Project. Here's What We Found. 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.

🚨 It's today.
AI on Your Own Terms: A Practical Guide to Local Models with Dillon Roach.

This is your sign to stop reading about local AI and learn how to build it.
✔ Free
✔ Practical
✔ Open source

🔗 Register & join → https://na2.hubs.ly/H05GLnW0

Welcome! You are invited to join a meeting: OSA Community event with Dillon Roach. After registering, you will receive a confirmation email about joining the meeting. Welcome! You are invited to join a meeting: OSA Community event with Dillon Roach. After registering, you will receive a confirmation email about joining the meeting.

Chuck McAndrew has a background that's hard to match. He came up through the US Marines, spent seven years doing infrastructure engineering in the public sector, and now leads infrastructure engineering at OpenTeams

He brings that same conviction to every engagement: technology exists to serve people, not the other way around.

On May 28th, he's leading our session on model provenance and federal audit readiness with Carahsoft.

If compliance and governance are on your plate, this one's for you.

Register → https://na2.hubs.ly/H05DSMW0

Tomorrow. Don't miss it.

Run a capable AI model on your own hardware, without a subscription, without sending data to the cloud, without giving up control!

Dillon Roach, will walk you through:

✅ Model selection & quantization (what actually runs on your machine)

✅ llama.cpp, vLLM, SGLang, demystified

✅ Connecting local models to code editors & tool-use frameworks

✅ OpenRouter & cloud GPU options when local hits its limits

This is an OSA Community event — built for senior architects and engineers who want the real picture, not the hype.

📌 Register now → https://na2.hubs.ly/H05DSv50



Running LLM agents locally gives teams more control over cost, data, and infrastructure.

But local ex*****on does not remove the hard parts of building reliable agents.

Adam Lewis, explored this by building a Harvest time-tracking agent with Pydantic AI, then switching the agent from Claude to a local model.

The experiment exposed a familiar AI challenge: a model can understand the request but still fail on the precise details that real workflows depend on.

This article shares what Adam learned about building agents that stay useful even when the model is imperfect. ➡️ https://na2.hubs.ly/H05Cwb00

One of our own is going through something no family should have to face alone.

Our colleague and friend, Logan McKee, suffered a hemorrhagic stroke while traveling for work in Rome on May 10th. He's been in critical condition ever since, with his wife by his side, and three young kids (8, 5, and 2) at home waiting for news.

The OpenTeams family is sharing this not to pressure anyone, but simply because showing up for people matters.

If you're moved to help, Logan's family has a GoFundMe here: https://na2.hubs.ly/H05BVBQ0

💙 Logan and his family are in our hearts.

Donate to McKee Family, organized by Haley Laughlin Our family is asking for support for my brother-in-law, Logan, who recently suffered a sudden and serious … Haley Laughlin needs your support for McKee Family