I work where technical constraints meet user behaviour — across SaaS, EdTech, Web3, and platform products — and turn the two into roadmaps that hit the number and ship on time.
Translating ambiguous business problems into prioritised, shippable roadmaps — and aligning stakeholders before a single sprint begins.
Taking products from problem to PRD to tested build — defining MVPs, pricing models, and the flows that actually convert users.
Moving the metrics that matter — screen time, retention, revenue — with data-backed experiments grounded in real user behaviour.
Shipping on time, every time — 100% on-time releases across 10+ sprints, zero escalations, and renewals trending up.
From EdTech at scale to enterprise SaaS to a hospitality platform built from zero — here's the work, and what it moved.
Built a hotel-management SaaS for independent Indian properties from zero — PMS, channel manager, booking engine, revenue management and AI Ads manager — with a pricing model backed by API-cost and margin analysis.
Turned ambiguous enterprise requirements across fintech, gaming, and DeFi into on-time technical roadmaps — aligning 20+ clients before kickoff to kill rework early.
Lifted average screen time +30% on a 1.2M+ subscriber product through data-backed UX optimisation — then carried those engagement and retention gains across the wider EdTech experience, managing 50+ escalations with zero brand incidents.
Now at India's largest overseas-careers & immigration consultancy — bringing B2B-SaaS and EdTech product experience to a high-volume B2C journey across visas, study-abroad, and job search.
I'm at my best where the problem is ambiguous and the execution has to be precise. If that's the work, I'd like to hear about it.
Currently based in Hyderabad, IndiaClearHost is a B2B SaaS hotel-management platform for independent properties in the Indian hospitality market — built from scratch, covering PMS, channel management, booking engine, revenue management, and AI Ads management. I led end-to-end product definition, including the PRD, and owned the go-to-market pricing strategy from the ground up.
Independent properties are price-sensitive, and the market already has established players — Aiosell, StayEzee, and others — competing largely on cost. But ClearHost's channel-management layer ran on Channex, a third-party API with real per-property licensing costs. That meant pricing wasn't just a competitive-positioning exercise; it was a margin problem. Price too low to win against incumbents, and the Channex cost alone could erode the margin on every account. Price without accounting for it at all, and the model breaks the moment it scales past a handful of pilot properties.
The tiers also weren't equivalent in cost to serve. A property on the base PMS tier had no Channex dependency at all; a property on channel management or full hosting did. A single flat price across tiers would have meant either overcharging the simplest tier or quietly subsidising the most expensive one.
I built a per-tier cost and margin model before finalising pricing, rather than setting prices first and checking margins after:
The model gave the team a tiered structure where margin held up across the board — none of the tiers thin enough to break under scale, none priced in a way that ignored the real cost of the channel-management dependency. The highest tier, which carried the most product surface area, also landed as the most attractive margin-to-price option. Annual billing discounts were set deliberately to improve cash-flow predictability while still protecting margin at every tier.
Pricing a 0→1 product against a competitive, cost-sensitive market only works if the cost side is modelled with the same rigour as the competitive side. It's easy to anchor purely on what competitors charge; the harder, more durable work is making sure each tier's price actually clears its own cost to serve — especially when that cost isn't uniform across tiers.
At The Matrix Labs, I owned end-to-end product roadmaps for 20+ enterprise clients across fintech, gaming, and DeFi — covering platform architecture, marketplace products, and audit tooling. Each client came in with a different level of technical maturity, and most arrived with business requirements that were directionally right but not yet scoped: “we need better fraud detection,” “make onboarding faster,” “support multi-chain.” Translating that into a sprint-ready roadmap was where the actual product work happened.
Early on, requirements were getting finalised close to sprint kickoff — sometimes during it. A stakeholder would flag a missing edge case, a compliance constraint, or a change in scope after engineering had already started building against the original spec. That meant mid-sprint rework: re-scoping tickets, renegotiating timelines, and absorbing the cost of decisions that should have been made upstream. It wasn't any one client's fault — it was a structural risk whenever ambiguous, cross-functional input had to be translated into a fixed two-week build cycle.
I moved requirement alignment earlier and made it explicit rather than assumed:
The mechanism here wasn't complicated — it was discipline about when alignment happens. Most rework doesn't come from bad requirements; it comes from requirements that were never fully surfaced before the team started building against them. Pulling that conversation earlier, even by a few days, is consistently cheaper than absorbing it mid-sprint.
At BYJU's, I worked across a subscriber base of 1.2M+ users, with a mandate spanning retention strategy, content-driven engagement, UX optimisation, and high-stakes escalation management. Retention in EdTech is a compounding problem — a subscriber who disengages early rarely re-engages on their own, and the cost of churn shows up both in revenue and in the broader health of the platform's usage metrics.
Retention efforts that aren't grounded in how users actually consume content tend to be generic — blanket reminders or one-size-fits-all engagement pushes that don't address why a specific subscriber is disengaging. The goal was to move away from that and build retention programmes that were data-backed: tied to actual content-consumption patterns rather than broad assumptions about what keeps users engaged.
I designed and executed retention programmes grounded in behavioural and consumption data, rather than treating retention as a single campaign-level lever:
Retention work holds up better when it's anchored to actual usage signals rather than generic engagement tactics — the targeting matters as much as the intervention itself. That principle carried through the rest of my work at BYJU's, including the UX-optimisation initiatives run alongside this.