AI Engineering Productivity, Code Reviews, and Domain Expertise
Jayeeta Putatunda · AI Engineering Lead · Turing
AI coding tools can generate more code, documentation, and prototypes, but that output creates new challenges for engineering teams. Jayeeta Putatunda, AI Engineering Lead at Turing, explains how AI changes code reviews, product handoffs, and expectations around engineering productivity. She also explores LLM observability, domain expertise, and why production systems still need rigorous engineering judgment.
Jayeeta describes the tradeoffs of AI assisted coding. Generated code and documentation can be verbose, complicating maintenance and review. As teams create more pull requests, review bandwidth can become a bottleneck. She discusses building review agents around an organization's rules so developers can focus on the most important changes.
Product teams and business stakeholders can now create prototypes, test ideas, and refine designs with fewer handoffs. That makes collaboration faster, but it also creates pressure to deliver production ready software just as quickly. Jayeeta explains why QA, permissions, testing across environments, and other engineering fundamentals cannot simply be skipped.
The conversation also covers LLM observability and the need to trace agent decisions, especially for sensitive applications such as finance. Jayeeta discusses how specialized models and domain experts can bring context to difficult workflows, and why engineering teams should share lessons to avoid duplicating work as AI tools change.