Tal Sheffer is a software engineer and technical writer who builds intelligent systems that help developers write better code. At Qodo, he focuses on retrieval-augmented generation (RAG) and agentic workflows, designing LLM-powered tools that give developers accurate, context-aware feedback right where they’re working.
His recent work includes building RAG pipelines that stay reliable even when the source material is fragmented or constantly changing, and integrating LLM agents into CI pipelines without disrupting existing workflows. Whether it’s writing logic for code review or tuning how models handle messy commit histories, Tal works close to the systems that engineers actually use day to day.
Before Qodo, Tal spent years shipping backend infrastructure and developer tooling across early-stage startups and larger teams. His code has powered low-latency APIs, internal dev environments, and everything in between, all with an eye for clean abstractions and maintainability.
Tal also writes about the systems he builds. His technical articles, including for NVIDIA’s developer blog, cover topics like LLM runtime optimization and system-level design for AI-assisted development. Tal also holds both a B.Sc. and M.Sc. degrees in computer engineering (summa cum laude) from Ben-Gurion University.
Colleagues know him as the engineer who is never afraid to try new approaches, innovate and take leadership in solving complex technical problems as well as a writer who explains things without oversimplifying.
