In this episode of Elixir Wizards, Charles Suggs and Emma Whamond are joined by Zach Daniel, creator of the Ash Framework and Igniter, VP of Engineering at Remedy Meds, and upcoming ElixirConf keynote speaker, to talk about what sits between an LLM and useful engineering work.
Zach reflects on how much has changed since his last appearance on the podcast in October 2024, moving from Igniter, code generation, and project patching into AI agents, context layers, and custom engineering workflows. The conversation explores how deterministic tools and probabilistic LLMs can work together, and why the most useful AI systems often depend on the structure built around the model.
We also discuss why teams should be careful about outsourcing the systems that hold their organizational knowledge and decision-making. He shares his perspective on owning the AI stack, building internal knowledge systems, training junior developers in an AI-augmented world, avoiding vendor lock-in, and why Elixir may be especially well-suited for safer agentic workflows.
Zach will be a keynote speaker at ElixirConf 2026, September 10–11 in Chicago, and the Elixir Wizards will be there too! Join us and the broader Elixir community, and use promo code Elixirwizards for 10% off in-person or virtual tickets at https://elixirconf.com/
Key topics discussed in this episode:
Zach Daniel’s work with Ash, Igniter, and AI tooling
How software development has changed since 2024
Deterministic code generation vs. LLM-generated code
Combining structured tools with AI agents
What it means to own your AI stack
Organizational knowledge as an engineering “spinal column”
Context layers, documentation, and internal workflows
Building custom agentic systems
Security, vendor lock-in, and open source LLMs
Junior developers and apprenticeship in the AI era
Podden och tillhörande omslagsbild på den här sidan tillhör
SmartLogic LLC. Innehållet i podden är skapat av SmartLogic LLC och inte av,
eller tillsammans med, Poddtoppen.
