SM
Product manager · builder · systems thinker

I build thesystems behindthe systems.

Product manager and builder who turns messy, high-friction operations into platforms that scale — from a recipe app I built solo to the AI knowledge tools and cloud infrastructure I've shipped for 100K+ users.

Say hello on LinkedIn View résumé
5+
years building productsAcross support operations, developer tooling, AI knowledge platforms, and cloud infrastructure.
$5.6M
in annual savingsDelivered by retiring a legacy platform across 30,000 pipelines and 6,000+ repos in under 6 months.
115K+
monthly usersOn an internal AI knowledge platform powering 300+ agents.
2.4×
answer qualityTook retrieval accuracy on that platform from 27% to 66%.

About

A systems thinker, with a founder's instinct.

I grew up in New Delhi, where at 16 I started and ran Just Football Co., a youth soccer company, through the rest of high school. It was the earliest evidence of a pattern: I like building things from scratch, and I've always known I want to found something of my own.

I moved to the US to study Industrial & Systems Engineering at Purdue, with minors in ECE and Entrepreneurship. Since then I've worked across support operations, developer productivity tooling, AI knowledge platforms, and cloud infrastructure. The common thread has never been the company; it's the type of problem: messy, high-friction operations that need a system, not a patch.

I think of product management as systems design more than feature design. The best work I've done wasn't a single launch—it was building the platform, data model, or operating process that let a dozen other launches happen faster. Outside the day job, that same instinct shows up in Kosh, a recipe app I built solo end to end.

What pulls me into a problem is simpler: does this make someone's life easier, and does it feel obvious once it's built? I care more about good design than most PMs are supposed to admit. If a system needs a training doc to be usable, it's not done yet.

I also spend time mentoring junior PMs, college students trying to figure out if product is the right path for them, and international students navigating a system that wasn't always built with them in mind—something I remember needing myself.

Outside work: cycling, golf, sailing, snowboarding, cooking, cocktails, and Formula 1. I speak English and Hindi natively, plus some Spanish and German.

How I think

A few operating principles.

06 principles

  1. Migrations need will, not just good tech.

    Migrations don't get killed by bad tech. They get killed by lack of will. I learned this retiring a legacy platform across 6,000+ repos and 30,000 pipelines. The technical plan was the easy part. The hard part was sequencing incentives so every team on the far end of that migration actually had a reason to move before their deadline, not after.

  2. An AI system is only as trustworthy as its worst wrong answer.

    Taking answer quality from 27% to 66% on an internal knowledge platform taught me that adoption isn't a retrieval problem, it's a trust problem, and trust breaks asymmetrically. One bad answer costs more goodwill than ten good ones earn back, which changes what you optimize for first.

  3. Cross-org alignment is a design problem, not a persuasion problem.

    Unlocking a cross-tenant data policy change required buy-in from stakeholders who had no shared incentive to say yes. The unlock wasn't a better pitch, it was reframing the ask so each side saw its own win in the outcome. I treat alignment as something you architect, not something you charm your way into.

  4. Systems should get quieter as they scale.

    Whether it's an onboarding flow, a support queue, or a capacity planning process, my instinct is to look for where the system is asking humans to do work a rule or a model could do instead. That's the throughline from automating support triage to automating demand intake now.

  5. Good design is the stuff people never have to think about.

    Customer obsession isn't a value on a wall for me, it's asking one more why than feels necessary until the actual friction shows up: the extra click, the unclear label, the workflow that only makes sense if you already know how it works. The best systems just work, and nobody notices they exist.

  6. I like building before I have permission to.

    I started a youth soccer company in New Delhi at 16 with no playbook and no one asking me to. That instinct never really left. It's why Kosh exists as a personal project, and why I spend time mentoring junior PMs, college students exploring product as a career, and international students navigating a system I once had to navigate myself, not because anyone assigned it, but because I noticed the gap.

Selected work

Things I've built and helped move forward.

06 projects

Experience

Building at scale, one system at a time.

Microsoft · 2020–present

  1. 2020

    PM Intern, Azure Infrastructure, Microsoft

    Built KPI frameworks and Power BI dashboards that cut bad rollouts by 70%.

  2. 2021–2023

    TPM, Azure Engineering Systems, Microsoft

    Led a three-platform consolidation, delivering $5.6M in annual savings and an Azure Impact Award in 2023.

  3. 2023–2024

    PM, Azure Developer Productivity & AI Tools, Microsoft

    Built AI support automation that cut MTTR from seven days to two and deflected 60% of tickets.

  4. 2024–2026

    PM, Azure Office of the CTO, AI Enablement, Microsoft

    Sole PM for an internal knowledge platform with 115K+ MAU, 1.15M+ articles, and answer quality improved from 27% to 66%.

  5. 2026–present

    PM II, Azure Capacity Management, Microsoft

    Leading global GPU and HPC capacity planning, with a focus on demand-intake automation.

Contact

Let's build something.

I'm always up for a conversation about AI platforms, knowledge systems, or the operational problems that show up once you're running things at real scale—whether that's inside a big company or building something new.