Writing
Technical deep-dives on distributed RL, agentic AI, and systems engineering.
State Machines for Structured Computation with Pydantic Graph
Modeling algorithms as explicit state machines over graphs instead of implicit control flow — using the Collatz sequence as a minimal example with pydantic-graph.
2026-04-305 min
From Intents to Actions: How Agentic AI Transforms Network Operations
How we combined LLM reasoning with reinforcement learning to build autonomous systems that translate high-level network intents into real-time control actions.
2025-02-1010 min
Scaling RL Training 20x: Lessons from Distributed Systems at Ericsson
Designing high-throughput distributed RL architectures with GPU learners, 100+ CPU actors, and multi-node coordination for radio network optimization.
2024-06-1812 min
When AI Has No Time to Think: Deploying Models Under Extreme Latency
Policy distillation and model compression techniques for deploying RL agents on baseband hardware with sub-millisecond inference budgets.
2024-03-228 min