Integrate AI into development processes in a large company (AI column)

I gave this talk for a month and a half at CTO Conf X, but the organizers decided not to publish the reports. So I decided on it. record for his channel And I'm going to talk about the state of the art approaches to how to decompose the work of engineers into jobs to be done scripts, and then how these scenarios can be improved. In the process, I am talking about both classical approaches to writing code, and options for helping with API design, conducting code review, running tests, eliminating vulnerabilities or detecting anomalies in the operation of solutions. The report will not only discuss SOTA, but also explain what really works for us and how we approach assessing the impact of these methods. In an hourly report, I discussed the following topics:

• Introduction and outline of the report The history of the emergence of AI copilots (Copilot and analogues)
• What is vibe coding? AI Prospects in Development, Anthropic’s MCP Protocol Agent interaction protocols and integration issues DevOps and Platform Transformations in Large Companies Developer experience and cognitive complexity reduction Measuring the goals and effectiveness of developers Google Governance API about AI-Enhanced API Design
• Code review in ByteDance Large scale migration using AI in Uber
• Booking experience (API, code review, migrations) Observability DataDog platform and AI On-call Engineer T-Bank’s approach to AI in SDLC AI-Copilot T-Bank Nester and its capabilities The Importance of AI for Large, Medium and Small Companies

By the way, I have already transcribed these reports in three parts: 1, 2 and 3

#AI #PlatformEngineering #Engineering #Software #Processes #Productivity