Ashish Desai

I'm a software engineer who got obsessed with AI before it was everywhere — specifically, with making it actually work in real products, not just in notebooks or demos.

Most of my work lives at the intersection of backend engineering and AI systems: agentic pipelines that use tool-use loops to reason and act, RAG systems that don't hallucinate, and APIs that can actually handle production load. I've built across the stack — Python and FastAPI on the backend, React Native and Expo for mobile, Next.js on the web.

I founded Solvbit because I kept seeing the same pattern: companies paying expensive consulting firms for AI strategies that no one could actually implement. Solvbit is the answer to that — a small, senior team that ships production-ready AI systems, not slide decks.

Outside of client work, I build things for myself. My current side project is a Claude-powered household inventory agent — it recognizes items from photos, tracks what you're running low on, and can tell you what to cook with what you have. Sounds small, but it's a great testbed for multi-step tool use and mobile AI integration.

I believe the best way to understand a technology is to build something with it that you'll actually use. I'm also convinced that most AI systems fail not because of the model, but because of the data, the infrastructure, or the product design around them.

If you're working on something interesting — an AI product, a tough data problem, a system that needs to scale — reach out. I'm always up for a good conversation.