Sowmya Yalavarthi

My journey has been a mixture of formal education and very hands-on learning through real work. I started in data engineering, where I learned how to work with large-scale data by building pipelines and making systems reliable. That experience shaped how I approach solutions in the generative AI field today, because I naturally think about production performance problems and how data flows through the entire system. I've recently moved into generative AI and started heavily working on LLM applications, which is the recent technology going on in the cloud. A big part of my growth also came from self-learning, building side hustles and personal projects, reading papers and docs, and experimenting with frameworks in current scenarios. My day-to-day involves building chatbots, testing them, and aligning them with the business. I start by checking how the system is performing in production, looking for any errors or staging issues, then do response quality checks followed by latency checks. I spend a good part of my day on the engineering side, working on RAG pipelines and turning them into chunks. I build and maintain the backend pieces using FastAPIs and integrations with vector databases. There's a lot of collaboration with stakeholders, business teams, and QA teams. I review edge cases with them and make sure the output is as they requested, so we're shipping something that solves real problems. At the end of the day, my responsibility is making sure the solution from the business is accurate and scalable for the real end users.