Sitemap
Medium Logo

I'm Yuval, an agentic AI specialist focused on the unglamorous side of AI agents—the part after the demo works and before the system survives contact with real users.

My technical home is multi-agent systems: LangGraph orchestration, memory architectures (RAG, MemGPT/Letta, Mem0, and Zep), LLM post-training (GRPO, DAPO, RLVR, DPO, and PPO), and inference infrastructure on Modal and vLLM. I spend most of my time thinking about how agent systems fail—and how to make them not.

On Medium, I write for engineers who are past the "hello world" stage and want the tradeoffs, the gotchas, and the production war stories that don't make it into the official docs. You'll find deep dives on reasoning patterns, chain-of-thought verification, multi-agent coordination, and the inference tricks that actually matter at scale.

Connect with Yuval Mehta
Yuval Mehta

Yuval Mehtahe/him

Agentic AI Specialist. I build multi-agent systems and write about what actually breaks in prod — LangGraph, memory, inference, post-training