Ben Labaschin · AI engineering & writing
Writing about how AI systems remember, fail, and scale.
Posts, talks, tools, and the work behind Agent Memory—now in Early Release from O’Reilly.
Agent Memory
Building Stateful AI Agents That Remember, Adapt, and Work Across Time
A practical engineering book about deciding what becomes memory, writing it safely, retrieving it well, and operating the system over time.
Current Focus
Building
Enterprise GenAI
Founding engineer at Workhelix, building the Nucleus platform that helps enterprises measure and grow AI ROI. Async LLM APIs, embedding pipelines, and agent deployment for Fortune 50 customers like Autodesk and Nasdaq.
Founding Engineer
Writing
Agent Memory for O’Reilly
Writing a practical guide to stateful AI agents that remember, adapt, and work across time. Chapters 1–3 are available now, with new chapters arriving throughout Early Release.
Early Release
Speaking
Agent Memory on Stage
Turning the book into practical talks about persistent state, context, and memory systems. Speaking throughout 2026 at conferences and meetups.
2026 engagements open
Related articles
Based on shared topics and references in the writing.
Selected article
Agent Memory Is in Early Release
The engineering questions behind Agent Memory, and how reader feedback can shape a working technical book.
Read articleWhat Are AI Agents? An Introduction to AI Agents and LLMs
Why relatedIntroduces the report “What Are AI Agents?” cited in the selected article.
Publishing for O'Reilly
Why relatedBoth cover O'Reilly, AI Agents, and Publishing.
Tools and Strategies for Agentic Development (Into the Hopper Podcast)
Why relatedBoth cover AI Agents.
Browse these 7 articles
- Agent Memory Is in Early ReleaseJuly 2026
- Publishing for O'ReillyNovember 2023
- Extending 'GPTs Are GPTs' to FirmsJune 2025