Reliable agent systems
Why agent orchestration, supervision, observability, and recovery matter after the demo works.
Speaking
I speak about reliable agent systems, Elixir, open source, and the leadership practices that help technical teams do ambitious work.
Topics I can speak about
Available for selected conferences, podcasts, panels, and engineering leadership sessions.
Elixir and the BEAM for AI
What actor-model systems and OTP can teach us about building concurrent, long-running agent workloads.
AI-first engineering teams
How teams can redesign work around AI while keeping human judgment, ownership, and useful measures.
People First leadership
Practical lessons on context, decentralized command, direct feedback, and Extreme Ownership.
Three decades of platform shifts
Lessons from open source, jQuery, startups, enterprise platforms, blockchain, Elixir, and AI agents.
Building open-source ecosystems
How to turn a useful technical idea into understandable tools, shared standards, and a contributor community.
Featured appearances
Recent conversations to listen to or watch.
Mike Hostetler and Jido: The Future of Elixir Agents
A conversation about the path to Elixir and the design of Jido as an autonomous agent framework.
Mike Hostetler on Autonomous Agents
A technical discussion of agent architecture, signals, supervision trees, and the future of agent-driven software.
Mike Hostetler on ReqLLM
A deep dive into model-provider differences, streaming, tool calls, structured output, and the Elixir AI ecosystem.
Earlier talks and community work
Selected entries from the archive.