I BUILD THE
INFRASTRUCTURE
THAT MAKES AI
ACTUALLY WORK.
Senior Software Engineer at Clari. I architect AI platforms, distributed systems, and LLM-powered agents. Built for scale, not just demos.
ABOUT
The story behind the systems.
Currently at Clari
Software Engineer · BLR, INDIA
“Any sufficiently ambitious system is indistinguishable from magic. Until someone builds it.”
WHO I AM
I think in systems. I reason in outcomes. The technical part is table stakes, what I actually care about is why a system needs to exist and what it needs to survive.
Somewhere along the way, I started asking the question founders ask. What happens when great engineering meets market instinct at the same time? That thread is still unspooling.
GET IN TOUCH →EXPERTISE
The areas where I operate at the edge.
AI Platforms & Orchestration
Production-grade AI platforms, from zero. The hard part isn't the model, it's the orchestration, fault tolerance, and making it debuggable at 2 AM. I've solved all three.
- Insights Extraction Engine
- Aggregation & Clustering Platform
- Topic Extraction Pipeline
- Agent Orchestration System
Distributed Systems
Built a custom orchestration system for data pipelines consumed by AI agents. No off-the-shelf tool fit precisely enough. I think in systems: data flow, bottlenecks, 10x load.
AI Agents & Coding Tools at Scale
Built LLM-powered agents and led AI tooling adoption across engineering: Augment, Claude Code, Cursor, Factory. From evaluation to org-wide standards. Engineers move 10x faster. See how below.
Java & Spring Boot
Robust, fault-tolerant RESTful services in Java and Spring Boot. Reduced the memory footprint of existing services by 40%.
First Principles Thinking
Every system, every problem, decomposed to first principles. What is this really trying to do? What's the simplest architecture that survives reality?
The question isn't whether to use AI coding tools.
It's how. And most orgs get this wrong.
I've led AI enablement across engineering teams, evaluating, selecting, standardising, and scaling coding assistant tooling from zero. Not as an experiment. As a force-multiplier that actually shipped.
MY ARSENAL
Whether you're just starting or already stuck mid-rollout.
WHAT THIS ACTUALLY LOOKS LIKE
Evaluate & Select
Ran structured evaluations of coding assistants across real engineering workflows, not benchmarks. Picked what actually moves the needle.
Standardise Practices
Built team-wide conventions: how to prompt, how to review AI output, when not to use it. The stuff no vendor tells you.
Drive Adoption at Scale
Went from sceptical engineers to org-wide dependence in weeks. The bottleneck was never the tool. It was the rollout strategy.
Measure & Iterate
Tracked real productivity signals, not vibes. Iterated on tooling choices and practices based on what the data said.
LET'S BUILD
SOMETHING
WORTH SCALING.
If you're building something ambitious in AI or infrastructure, or just have a hard problem to work through, I'm genuinely interested to help!
- →AI Platform & Infrastructure Advisory
- →Technical advisory for early-stage founders
- →Hard engineering problems at scale
SEND A MESSAGE