Who am I?
I'm a backend and AI-systems engineer with a soft spot for clean architecture and fast, boring-in-a-good-way infrastructure. I graduated from IIT Roorkee with a B.Tech in Materials Engineering, but software was always where my curiosity lived.
I currently build Clara — the AI Source-to-Pay Agent at Procol — where I work on event-driven integrations, MCP platforms, agent memory, and the kind of performance tuning that shows up on the bill. I like the parts of a system nobody sees but everyone depends on.
Most of what I know came from side projects, production fires, and reading code. I build to understand, then I build it again better.

Where I attended college
I pursued my B.Tech in Materials Engineering from the Indian Institute of Technology Roorkee — one of India's oldest and most prestigious engineering institutions.
While my degree was in Materials Engineering, I discovered my passion for software development during my time here. I spent countless nights building side projects, contributing to open source, and teaching myself full-stack development.
IIT Roorkee gave me the discipline and problem-solving framework that I carry into my work today. It's where I learned that engineering thinking transcends any single domain — and that the best solutions come from first-principles reasoning.
Indian Institute of Technology Roorkee

Where I am working
I'm currently working as a Senior Software Engineer at Procol, where I'm building Clara — the AI Source-to-Pay Agent that's transforming enterprise procurement.
I architected Clara's event-driven system from scratch, enabling multi-platform integrations with Procol, Ariba, Coupa, and SAP. I built the MCP Integration Platform that dramatically reduced integration time for enterprise clients, and designed the Long Term Memory Store that gives Clara episodic memory across sessions.
Beyond Clara, I work on optimizing our infrastructure — recently reducing server costs by 25% while more than doubling our throughput capacity. Every day brings new challenges in distributed systems, AI service orchestration, and large-scale event processing.

Where I worked
A few places and people that shaped my journey. Each experience taught me something new about building great software.
Enhanced analytics services reducing response time by 40%. Improved sourcing module performance with 80% latency reduction on 3 main APIs. Built Integration Studio from scratch, leading to 3x more integrated clients.
Led Rails 5 to Rails 7 migration with minimal downtime. Developed a PDF rendering service for document generation. Contributed to main backend and frontend services.
Worked on OCR algorithm improving efficiency by 20%. Implemented CI/CD pipelines to streamline deployment and reduce time to market.
Built a full-stack dating app helping users find trustworthy singles through referrals. Handled everything from architecture to deployment.
Beyond code
Code is what I do, but not all of who I am.
Basketball & NBA analytics
I watch games like a systems problem — roster construction, lineup fit, coaching adjustments, and the small decisions that decide close series. Basketball is maybe the best place outside of software to see compounding edge cases in real time.
Guitar
I play to unplug and reset. There's something about muscle memory and timing that translates surprisingly well back to debugging patience.
Reading
Harry Potter and Marvel on repeat, long tech articles, Hacker News rabbit holes, and the occasional audiobook. I read to find surprising ideas in unexpected places.
Geopolitics
Foreign policy, international relations, and how history, economics, and incentives collide. It keeps the part of my brain that likes tradeoffs busy.
Software archaeology
Digging through git history, old commits, PR threads, and dependency changelogs to understand why code exists. The best explanations are usually buried, not obvious.
“The best engineers I know bring something non-engineering into how they think. For me that's basketball lineups, guitar practice, and the occasional deep dive into why a 2018 dependency choice still matters.