Talks
Talks on AI engineering, large codebases, production readiness, and choosing a stack.
Frameworks are not enough: The future stack for AI development
AI makes code generation dramatically faster, but faster code does not automatically mean better software.
Given at FrontKon 2026, Slides
AI Does Not Remember. It Only Stores.
Most AI products call it memory, but it is only a database.
Given at Černá AI Festival 2026, Slides
Local or Cloud Models?
A practical comparison of local, on-premise, and cloud AI models: hardware, formats, runtimes, context, parallelism, and hybrid deployment.
Given at AI Monday Jihlava #7, Slides
Vibe Coding vs Agentic Engineering
Where hype-driven generation ends and production-ready engineering begins.
There Is No Best Stack
Technical choices only work when they fit the team, codebase, and constraints.
AI Engineering on Large Codebases
How project memory, standards, code review, and documentation keep AI useful on real codebases.
Production-Ready AI Engineering
Quality gates, ADRs, observability, and coding agents teams can trust.
I adapt each talk to your audience, length and format.