Skip to Content
View all events

AI Superstream: AI Harnesses

Published by O'Reilly Media, Inc.

Beginner to advanced content levelBeginner to advanced

Build the frameworks that turn models into agents

Forget AI-model leaderboard hype—are you building the harness?

The real breakthrough is the invisible infrastructure that transforms standalone LLMs into dependable agentic systems. By managing tools, memory, planning, and self-correction, the harness provides the guardrails necessary for true autonomy. Learn why harness engineering now matters more than model size and how this architecture is shaping the future of AI.

Join our experts at this live event designed for AI engineers and agent builders and discover the powerful frameworks required to ship reliable, production-ready agents.

We’re still working on finalizing the schedule for this event. Please check back closer to the event date for more information.

What you’ll learn and how you can apply it:

  • Understand the role of the agent harness and how it enables transparent, auditable, and correctable AI systems
  • Explore how procedural memory helps agents learn from past actions and improve performance over time
  • Learn how harness design can improve retrieval and reasoning workflows, helping smaller models achieve stronger results at lower cost
  • Design agent systems that coordinate planning, memory, and feedback loops to improve reliability and control

Recommended follow-up:

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

Introduction – Christina Stathopoulos (5 minutes)

Christina welcomes you to the AI Superstream.

Build Your Own Agent: Escaping the Harness – John Berryman (45 minutes)

“Agent harness” has become the default term for tools like Claude Code, Cursor, and Codex, but that term obscures the fact that the agent harness is the agent, and you can build one yourself. John Berryman, founder of Arcturus Labs, constructs a simple agent from first principles, showing just how approachable it really is. You’ll see how writing agent skills in plain English makes the agent’s behavior transparent, auditable, and correctable by you and by your customers. And he’ll discuss what’s on the horizon: agents that aren’t trapped in the terminal, that can be wired into any application, and that carry personal context wherever they go.

Break (5 minutes)

Procedural Memory Is All You Need: The Continuous-Learning Primitive Your Agent Harness Is Missing – Casius Lee (35 minutes)

Building an agent is easy. Keeping one alive in production isn’t. The layer that Casius Lee keeps coming back to is procedural memory, the part of an agent harness that lets it learn the job by doing the job rather than relearning it from scratch (and burning tokens) every time it runs. The Oracle AI developer advocate shares three failure modes he sees in production agent harnesses and discusses why procedural memory is the primitive that decides whether your harness survives in production.

Building Agent Harnesses to Match Frontier-Model Performance in RAG – Vineeth Kalluru (35 minutes)

Most enterprise RAG systems today plateau at retrieving documents and handing them to the largest model available, relying heavily on the model’s context window and system prompt to handle the rest. With recent advances in harness engineering, it’s possible to redesign RAG so that the harness manages search state, curates evidence, and verifies answers before leaving the model to do what it does best: generation. Vineeth Kalluru, generative AI solutions architect at NVIDIA, traces this evolution from single-shot retrieval to recent stateful, agentic search approaches, showing how the right harness lets a compact open model match frontier-model quality at a fraction of the cost. His talk includes a live hands-on tutorial based on a recently published research showcasing the concept.

Break (5 minutes)

The Agent Harness Interface: Code, Control, and Adaptive Systems – Nicole Koenigstein (35 minutes)

Once code is placed inside the agent loop, the harness has to do more than just wrap a model. It has to decide what to execute next, preserve useful state, expose the right tools, interpret feedback, and turn failures into corrective actions. That’s where agent systems become executable, verifiable, and stateful rather than just generative. Nicole Koenigstein, author and AI researcher, connects the idea of code as agent harness with adaptive agent systems. The focus is not on harnesses that rewrite themselves but on the interface through which agent behavior becomes observable and adjustable over time, from planning, memory, and tool use to feedback-driven control, model selection and effort, skill use, role assignment, and topology.

Session to Come (35 minutes)

Please check back for more information.

Break (5 minutes)

Session to Come (35 minutes)

Please check back for more information.

Closing Remarks – Christina Stathopoulos (5 minutes)

Christina closes out today’s event.

Your Hosts and Selected Speakers

  • Christina Stathopoulos

    Christina Stathopoulos is a global keynote speaker, award-winning educator, and the founder of Dare to Data on a mission to help individuals and corporations take the next step in their data and AI journey. After building a successful career at companies like Google and Waze, she shifted her focus to scaling impact through education, training, and product evangelism. Today, she’s trusted by an extensive client list of Fortune 500 and big tech organizations. Outside of her corporate work, Christina holds an adjunct faculty position at several universities, where she lectures and leads programs on data and AI for business. She’s also an outspoken ambassador for responsible AI and a voracious reader.

  • John Berryman

    John Berryman is the founder and principal consultant of Arcturus Labs, where he specializes in LLM application development and helps businesses harness the power of advanced AI technologies. As an engineer working at the forefront of AI-assisted coding tools, John was an early contributor to the development of GitHub Copilot’s completions and chat functionalities.

  • Nicole Koenigstein

    Nicole Koenigstein is an AI researcher and practitioner in agentic systems, working across research, consulting, teaching, and direct system implementation to build reliable, production-ready AI systems. Her work focuses on multi-agent architectures, evaluation, safety, and long-term system behavior. She’s the author of Math for Machine Learning and Transformers in Action (Manning Publications). Her forthcoming books, Transformers: The Definitive Guide and Applications Beyond NLP and AI Agents: The Definitive Guide, will be published by O’Reilly Media. Nicole served as an external evaluator for a European Commission AI Grand Challenge and has advised IOSCO on generative AI in regulated environments. She also serves on advisory boards for leading AI and quantitative finance conferences and regularly delivers invited talks and technical workshops across academia, industry, and international events.

  • Vineeth Kalluru

    Vineeth Kalluru is a generative AI solutions architect at NVIDIA, where he helps enterprises deploy large-scale generative AI in production, across silicon, compilers, foundation models, and GPU inference. He has shipped open source data curation pipelines adopted by major AI labs and authored research on agentic skill optimization. Vineeth has spent his career tracing AI systems from the chip up to the application layer. Previously, he worked on developing MLIR-based compilers at SambaNova Systems and analyzing TPU performance at Google.

Skill covered

Artificial Intelligence (AI)