What I bring
to the table.
A full-stack PM skillset — from writing prompts for LLM agents to defining data models, from running 40+ user interviews to owning a ₹40 Cr product roadmap.
"A voice agent isn't a technical project — it's a speed-of-outreach problem." I start with the system constraint, not the feature list. Every decision is traced back to a metric. Every product is designed to compound.
The fundamentals,
with proof behind them.
Every PM claims these skills. The difference is in the decisions they've backed — from PRDs that went straight to dev, to OKRs that moved the business needle.
"I don't ship features that can't be traced to a metric. I don't start with solutions."
Not AI-aware —
AI-native.
Built 4 AI products in production — voice agents, RAG systems, LLM scoring engines, dynamic rendering. Each with guardrails, fallbacks, and real revenue impact.
Semantic deviation threshold, confidence gating, and graceful fallback — because AI in a revenue path needs hard constraints.
CAC down.
Revenue up. Every time.
Growth work that compounds — referral systems, affiliate engines, PLG loops, and landing page personalisation. Every project has a before/after in the numbers.
4 growth products. CAC reduced 35–65%. ₹40+ Cr revenue influenced. One channel built from zero to India's 2nd largest.
Deep enough to
not need a translator.
I write specs that go straight to dev. WebSocket relay architecture, Redis caching strategies, database schemas, API contracts — handed to engineers without a meeting.
Built Hercules ERP on Node.js. Defined ANIKA's audio pipeline at 24kHz. Wrote ECHO's semantic formula.
The toolkit I
actually use.
Not a tools collector. Each tool here has shipped something — a PRD, a spec, a dashboard, or a key product decision.
Microsoft Clarity for ECHO discovery. Mixpanel for emConnect funnel. Figma for every product wireframe.
Not in the playbook.
In the decisions.
Six frameworks that shaped every product — not theory, tools I reach for when the problem is hard and the answer isn't obvious.
First Principles Thinking
Strip the problem to its root. A voice agent isn't a tech project — it's a speed-of-outreach problem. Solved admissions leakage at scale.
Jobs To Be Done (JTBD)
Every product failure traces back to a JTBD mismatch. Used to define Emerge's 5-failure RCA and rebuild 42.ai around school incentives.
Systems Thinking
Features are nodes in a system. AURA: audit → intent → pitchmate → outcomes → back to audit. Each layer amplifies the next.
RICE Prioritisation
Reach × Impact × Confidence / Effort. Keeps roadmap decisions defensible in every stakeholder conversation without gut feeling.
Root Cause Analysis (5 Whys)
Before any pivot, a structured RCA. The Emerge → 42.ai pivot was driven by a 5-failure data analysis, not pressure or emotion.
PLG Loop Design
Growth is a by-product of value creation. Encoded seed → refer → earn → compound into emConnect — 44% joined via in-product referrals.
What I
don't do.
Self-awareness is part of the skillset. These aren't weaknesses — they're guardrails built into how I work.