Open Source AI Package of the Week
Published by O'Reilly Media, Inc.
Discover and decode the tools shaping AI development
New open source AI tools are emerging faster than most developers have time to evaluate them. Open Source AI Package of the Week helps you keep pace, with a fast, hands-on look at one noteworthy package, library, framework, or developer tool each week. You’ll see what each does and where it fits in the modern AI stack, and watch an expert put it to work in a practical live demo. You’ll leave knowing not only how to use it but when to use it, where it falls short, and whether it deserves a place in your AI toolkit.
What you’ll learn and how you can apply it
- Evaluate emerging AI tools and quickly understand what they do, where they fit, and what problems they solve
- Make smarter tooling decisions by understanding each tool’s strengths, limitations, and best use cases
- See the tools at work in live demos and practical examples and understand how to apply them in your own AI projects
This live event is for you because...
- You’re an AI/ML engineer, developer, data scientist, or technical practitioner looking to keep up with the rapidly evolving open source AI ecosystem.
- You want a fast, hands-on way to discover and evaluate new AI tools before deciding whether they belong in your toolkit.
Prerequisites
- Come with your questions
- Have a pen and paper handy to capture notes, insights, and inspiration
Recommended follow-up:
- Read Large Language Models: The Hard Parts (book)
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Thursdays 8:00am PT / 11:00am ET
- Interactive discussion and Q&A (45 minutes)
Upcoming Sessions
November 5, 2026 – Bruce Hopkins
Get Started Using the Pi Agent Harness with Kimi K3
The Pi agent harness is a powerful but extremely lightweight open source package that developers can customize for their own needs. It doesn’t have out-of-the-box support for the Model Context Protocol (MCP), the predominant protocol for connecting AI agent clients and servers. This first session of Open Source AI Package of the Week will get you started with Pi. Bruce Hopkins shows you how to configure it with the Kimi K3 model and add support that allows Pi to connect to any MCP server.
November 12, 2026 – Ben O’Mahony
Pydanic AI is the Only Harness You Need
Pydantic AI is a type-safe Python framework for building, testing, and running AI agents across different models and providers. In this live-coding session, Ben O'Mahony shows you how the Pydantic AI harness makes it easy to build and extend AI agents. Starting with a simple agent, he’ll rapidly add new capabilities and make the harness self-improving. Along the way, you’ll see how Pydantic and a small amount of Python can support surprisingly sophisticated agent behavior.
November 19, 2026 – Kiruthika Subramani
Document Vision and Parsing Gauntlet
Are you paying a cloud vision API to read receipts that a free model could handle? Kiruthika Subramani puts Qwen 3 VL through a gauntlet of real paperwork including crumpled receipts, invoices, screenshots, and handwriting, and keeps score live on screen. Images go in, structured JSON comes out, and you’ll see exactly what it nails and what it confidently gets wrong. Find out whether an open model can take over your document work, and leave with the extraction pipeline to test it on your own paper.
December 3, 2026 – Richmond Alake
Inside DeepSeek Harness
What does an AI agent need beyond the model? DeepSeek Harness, an open source agent runtime, answers with one unusual design decision: Everything is a plugin, including the agent loop itself. Richmond Alake uses DeepSeek Harness’s architecture to walk through the components of a complete agent harness: model adapters, tools, context management, persistence and orchestration. At the center is the append-only session log, a durable record of every message, tool call, and state change. See how that event-driven foundation makes agent behavior easier to extend, recover, replay, and reason about.
Your Hosts and Guests
Bruce Hopkins
Bruce Hopkins is a widely-recognized AI expert. He's a Claude Certified Developer, an Oracle ACE and Intel Software Innovator (both in the fields of AI), as well as a Java Champion. He’s also the author of ChatGPT for Java and the coauthor of Beginning ChatGPT for Python.
Ben O'Mahony
Ben O’Mahony is a results-driven AI and engineering leader with a track record of building high-performing teams and shipping business-critical AI, ML, and data platforms at scale. He has deep expertise across the full engineering and data lifecycle, from research to production deployment and is adept at defining technical strategy, driving execution, and partnering cross-functionally to deliver measurable impact. He’s currently writing Building AI Agent Platforms for O’Reilly, due to be published next spring, and available to read now in early release on the O’Reilly platform.
Kiruthika Subramani
Kiruthika Subramani is a data scientist at Bell Canada who has authored two books on artificial intelligence, delivered more than 230 talks worldwide, and published over 100 articles and blog posts. She’s a Google Developer Expert in AI, an IBM Champion for Data and AI, a Women Techmakers Ambassador, and a 2025 Women in AI Scholarship Award winner. Her previous experience includes serving as head of AI at Musitechnic Formation and interning at Amazon. Kiruthika holds four cloud certifications from Google Cloud and five from AWS. She earned her master’s degree from Mila, an artificial intelligence research institute in Montreal.
Richmond Alake
Richmond Alake is director of AI developer experience at Oracle, where he leads developer relations for the Oracle AI Database and pioneers the discipline of memory engineering, a term he coined to describe the structured design of agent memory systems. He runs the #100DaysOfAgentMemory LinkedIn series, which has reached tens of thousands of AI practitioners.
Richmond has published over 200 technical articles with more than one million views across Medium, NVIDIA’s technical blog, and other industry publications, covering topics from agent memory systems to production LLM application development. He has spoken at the AI Engineer World’s Fair, DeveloperWeek, Confluence, and Monki Gras. Previously, he served as an AI/ML developer advocate at MongoDB, where he built MemoRizz, an open source Python agent memory library, and codeveloped a DeepLearning.AI course on prompt compression and query optimization. Richmond holds a master’s degree in computer vision, machine learning, and robotics from the University of Surrey.