Alok Gudikote

I've shipped code that 30 million people use every day — and then automated away 20% of my own team's code review time using AI. For 12+ years I've owned Android features end-to-end: architecture, build, production. At BYJU'S, I helped scale our events module to 20-30M peak daily active users during the remote-learning surge, while steering the shift from MVP to MVVM/Jetpack Compose and cutting crash rates from ~8% to under 0.75%. At Walmart Global Tech, I owned retention features inside the Walmart+ membership app — Activity Tracker, Savings Dashboard, Value Projector, Next Best Action — built with 95% test coverage and full TalkBack/ADA accessibility. I got into Android because I wanted to build things regular people would actually use, not just internal tools. That instinct never left. I've shipped a dozen+ personal apps to the Play Store, from a savings game to a programming-practice application. It's also where my best ideas at work come from — the AI-driven Mermaid diagram tool I built for pull requests started as a weekend experiment. Right now I'm focused on where Android architecture meets AI tooling: agentic workflows, AI-assisted media creation, and developer-productivity systems at scale. On the stack side, that means Python and FastAPI on the backend, PostgreSQL/pgvector for retrieval, LangGraph for agentic workflows, and Gemini, Claude, and OpenAI APIs for the AI layer — deployed on Vertex AI and Cloud Run. If your team is migrating a legacy stack to Compose, or figuring out how AI fits into an Android codebase beyond autocomplete, I'd like to hear how you're thinking about it.