Curiosity first. Code came after.
Rohit has always been drawn to the moment something starts to understand — a system that goes from following instructions to actually reasoning about a problem. That curiosity pulled him from electrical engineering into software, and eventually into a front-row seat for the biggest shift the field has seen: language models that read, retrieve and act. What keeps him hooked isn't the novelty of AI — it's the craft of making it reliable at enterprise scale: watching a system he built help a colleague find an answer buried in thousands of technical documents in seconds, or a translation tool bridge a language gap for hundreds of engineers on a rig floor. That same instinct for communication and storytelling shows up outside work too, on stage in inter-college drama competitions — a reminder that the best engineering, like the best story, is really about making something complex easy for someone else to understand.
Understand the Problem
Translate an ambiguous business ask into a concrete, solvable AI problem before writing a line of code.
Architect the System
Decide how data, models, retrieval and infrastructure fit together end-to-end, with ownership of the full design.
Build & Evaluate
Engineer with LLMs, RAG and agentic workflows — then benchmark rigorously before anything reaches production.
Ship & Operate
Deploy on cloud infrastructure with CI/CD and monitoring, so the system stays reliable long after launch.