Integrating an LLM into an Agent Without Frameworks
Extend a deterministic Python agent with LLM-driven decisions using the OpenAI API, structured outputs, state management, decision validation, and application-controlled execution.
Recent thoughts and experiments.
Technical deep dives, practical implementations and lessons from real projects.
Dynamic Decisions Create New Engineering Responsibilities
How dynamic action selection introduces new responsibilities for execution control, security, observability, and architectural decisions in AI agent systems.
Building an Agent Without Frameworks
Build a deterministic execution loop in Python from scratch, exploring execution state, termination conditions, failure handling, and the foundations of AI agent control.
An engineer exploring the intersection of software, AI and mathematics.
I'm a Senior Software Engineer with a strong interest in Artificial Intelligence, Computer Vision and the mathematics behind these technologies. Here I share technical insights, experiments and lessons from real projects.
Learn more about meKeep learning. Keep building.
Technical insights, practical experiments and engineering ideas. Every two weeks.
No spam. Just valuable content.
By subscribing, you agree to receive the Engineering Playbook and to our Privacy Policy.