Journey
I chose this path on purpose.
I’m an AI engineer on the founding team of aiRA, Capillary’s agentic AI platform. I didn’t arrive here by default. I studied Chemical Engineering at IIT Gandhinagar — a non-circuit branch — routed myself toward AI deliberately, then walked into industry as a product analyst and taught myself into engineering, earning a promotion each year.
I think of a career less as a sequence of jobs and more as the slow accumulation of engineering judgment. Most of what’s below was chosen, or grabbed, for what it would teach me.
The path
- 2020 – 2024B.Tech, Chemical Engineering — IIT GandhinagarMinors in Computer Science & Artificial Intelligence
A non-circuit branch, and a deliberate route into computing. I took the CS and AI minors and pointed my electives, research, and internships at how intelligent systems actually get built.
- 2022 – 2023Early proof — ML & data-science internshipsInfyU Labs · PUMA
Fruit-quality QA models at an agri-tech startup, then e-commerce demand forecasting at PUMA (LSTM with Bayesian optimization), which earned a pre-placement offer. First time watching models meet real operations.
- Jan – Jun 2024Research — LLM code-similarityIIT Gandhinagar · FSE 2024
Probed how robust code-generating LLMs are to prompt mutations. Accepted as a poster at ACM FSE 2024. Where my interest in how these models actually behave got serious.
- Feb – Jul 2024SDE Intern — CapillaryCapillary Technologies
Built a RAG Slackbot for the product team that 45+ people came to rely on. The internship that turned into a decision.
- Aug 2024Joined the aiRA founding team — as a Product AnalystCapillary Technologies
One of the first ~5 people on aiRA, Capillary’s agentic AI platform, at proof-of-concept stage. I came in on the product side and taught myself into the engineering.
- 2025Promoted to SDE-1aiRA · Capillary
The analyst-to-engineer move, made official — and the start of owning systems, not just features.
- 2026Promoted to SDE-2aiRA · Capillary
Feature → system → platform. Now working on turning aiRA from Capillary-bound into an orchestration layer across a customer’s stack.
How I think about engineering
-
Problems before technologies
The question is never “how do I use the newest framework?” It’s “what’s the simplest system that solves this well?” Sometimes that’s an LLM. Sometimes it’s SQL. Sometimes it’s deleting code.
-
Own it end to end
A feature isn’t finished when it compiles. It’s finished when it survives production and the next engineer can understand it. Ownership is a mindset, not an assignment.
-
Reason from first principles
I’d rather understand why a system exists, what constrained it, and what breaks it than copy an architecture because it’s popular.
-
Optimize for impact, not output
Lines of code aren’t achievements. Problems solved are. Every decision should answer: what measurably got better because this exists?
-
Build systems, not features
Features disappear; systems compound. The best work leaves the next problem easier to solve.
-
Simplicity is the advantage
Clarity over cleverness. Simple usually costs more understanding up front and far less pain forever.
Beyond the work
Outside engineering, I’m an athlete — or I keep insisting on being one. I played football through school and for IIT Gandhinagar’s inter-IIT team. I’ve since reconstructed the ACL in both knees, on different legs, about two years apart: the first in November 2023, the second — with a meniscus tear — in September 2025. Between the two surgeries I ran three half-marathons, including the Tuffman Bengaluru and the Apollo Delhi, and had started training for a full one. I’m back in rehab now, working a return-to-sport protocol toward playing again.
Rehab has taught me more about patience and honest progress than almost anything else. You can’t rush a healing ligament and you can’t fake the milestones — and it’s not lost on me how much that resembles the engineering I care about: long games, compounding effort, and no shortcuts that don’t eventually cost you.
I also sing Indian classical music and play the keyboard. It’s the part of the day that isn’t optimizing for anything, and I’d like to keep it that way.
Photos from the pitch and the road — coming soon.