Scaling agentic architectures in regulated industries
Keynote at Digital Day, Belgrade. How regulated industries can ship agentic AI without giving up compliance, control, or security.
Writing · Talks · Stickers
Separating the language from the execution.
Essays on AI agents and the unglamorous work of building software that doesn't embarrass you in production.
Doing chatbots since before they were good.
From NLP to LLMs: the quest for a reliable chatbot
On the a16z Podcast with Martin Casado. Why LLMs dull the sharp corners of natural-language interactions, and how to pair them with business logic that won't hallucinate.
Inside Rasa: navigating the future of conversational AI
On Bluejay's Skywatch. Rasa's evolution as open-source infrastructure, building reliable real-world conversational systems, and the rise of voice.
Beyond supervised learning: the rise of in-context learning with LLMs
On Vanishing Gradients with Hugo Bowne-Anderson. Where supervised learning failed, and how LLMs are reshaping the agent stack.
I'm so glad we got rid of intents.