Repeat entrepreneur working at the intersection of AI, fintech and consumer products. Co-founded Tipplr, one of India's fastest-growing ONDC food networks, and led it through a successful acquisition. Now building a portfolio of AI-first products through Nexus AI / CIQ AI Solutions — most of them live, links below.
work
building now
01 / askgogo
AskGogo
Your personal AI agent — it remembers, plans, acts, and keeps working when you're away.
Remind me about Dad's insurance renewal next month.
Got it. I'll keep it in memory and remind you before it's due.
REMEMBERPLANACTWATCH
Most assistants answer questions. Gogo gets things done. You tell it once, on WhatsApp or on the web, and it remembers. Hand it a task and it runs in the background, asks for your approval before anything irreversible, and reports back in the chat you already live in.
It's built for India first: voice notes in your own language, reminders that actually arrive, and your email, calendar and documents in one place. Gogo is evolving into a true personal agent — with persistent memory, a private browser per user, approval-controlled actions and background runs that keep working after you leave the chat. The waitlist is open.
WhatsApp is one way to talk to Gogo, not the product itself. The same memory and context are designed to follow you across chat, web and mobile surfaces.
What it is becoming
Persistent memory across conversations and documents
Your card points, turned into real flights and hotel stays — with honest math on what every point is actually worth.
CreditIQ turns reward points into usable travel decisions. A Fly on Points board and stay-on-points hotel view show what your existing balances can actually unlock across flights and hotels.
The redemption engine compares transferring points to airline or hotel programmes against using the bank's own portal, and keeps verified data visually separate from estimates. The principle is simple: show the real conversion path, the real constraints, and the real value before calling something a good redemption.
For HDFC SmartBuy, a Regalia point is worth ₹0.50 and points can cover at most 70% of a booking. Transferred points do not carry that same booking cap. CreditIQ makes differences like this explicit instead of hiding them behind headline reward rates.
Outcomes
Fly on Points board for real flight redemption options
Stay-on-points hotel intelligence
Transfer-partner versus bank-portal comparison
Verified figures shown differently from modelled estimates
Statement-based analysis to replace assumptions with actual spend
One of India's fastest-growing ONDC food networks — built, scaled and successfully acquired.
daily orders · 30 days
300 → 2,000
orders delivered
500,000+
cities
BUILT → SCALED → PROFITABLE → ACQUIRED
India's restaurant ecosystem was fragmented and small restaurants had little direct relationship with their customers. ONDC was the opening, and we built Tipplr to be one of the fastest movers on it.
We went from 300 to 2,000 daily orders in 30 days, delivered 500,000+ orders, onboarded 20,000+ restaurants across 10+ cities and grew revenue 350% in six months. The platform reached profitability and was acquired.
The stack was built for scale from day one — real-time order routing, dynamic pricing, inventory sync with restaurant POS systems, rider allocation, and a merchant dashboard that gave restaurants visibility they'd never had.
Outcomes
300 → 2,000 daily orders in 30 days
500,000+ orders delivered across India
20,000+ restaurants onboarded across 10+ cities
350% revenue growth in six months
Reached profitability, then a successful acquisition
Acquired
04 / rasoi capital
Rasoi Capital
AI-powered lending for HORECA — from underwriting to collections.
Hotels, restaurants and cafés are the backbone of Indian food commerce, and among the worst served by formal credit. Seasonality, daily settlements and informal bookkeeping make standard underwriting models fail.
Rasoi Capital's underwriting engine runs a six-factor credit model built for HORECA cash flow. It ingests POS data, GST filings, rental agreements and utility bills to build a risk profile traditional lenders can't see, and explains every decision in plain English.
The platform covers the full loan lifecycle: application intake, AI scoring, document vault, credit manager workflow, disbursement, active loan tracking with DPD buckets, a collections engine with field agent beats, EMI restructuring and legal notice tracking.
Outcomes
Six-factor credit scoring tailored to HORECA cash flow patterns
Risk narrative explaining every underwriting decision
Full lifecycle — application → disbursement → collections → legal
Portfolio analytics with live AUM, NPA rate and city breakdown
Collections engine with DPD dashboard and field agent assignment
Mutual fund portfolio analytics — AI signals, tax harvesting and rebalancing in one view.
Most Indian investors have no idea what their portfolio is actually doing. A CAS statement arrives every month — a dense PDF with hundreds of transactions across AMCs — and there's no easy way to make sense of it.
Upload the statement and FolioKey parses it in your browser. No passwords, no OTPs, nothing leaves your machine. The engine computes real after-tax returns, surfaces concentration risk, finds tax-loss harvesting opportunities and proposes rebalancing trades that respect exit loads, lock-ins and LTCG brackets.
It came out of a personal frustration — 11 funds across 4 apps, with no way to tell whether I was diversified or just duplicated. The first version parsed a CAS in under 30 seconds. The current one runs a portfolio health score, goal-based planning and a chat that answers "should I exit this fund?" with the data behind it.
Outcomes
Portfolio parses in-browser with zero server-side storage
Tax-loss harvesting engine surfaces realisable savings per portfolio
Government schemes, made findable — eligibility, documents and application tracking for Indian citizens.
India runs thousands of central and state welfare schemes. Most citizens who qualify never apply — they don't know the scheme exists, can't work out whether they're eligible, or can't assemble the paperwork it asks for. The barrier isn't the benefit, it's the distance to it.
SaralSetu closes that distance. A six-step eligibility check covers Karnataka and central schemes in English or Kannada, then walks the citizen through document preparation, application assistance and progress tracking — with a WhatsApp channel for people who will never fill in a web form.
The hardest design constraint was trust. A platform that helps with government paperwork gets mistaken for the government, so the product states on every page that it is an independent facilitator and that final eligibility and approval decisions rest with the relevant authority. Privacy policy, terms and a data-deletion route are written against the DPDP Act 2023.
Built and operated by SetuNexa Technologies — a separate company from my Nexus AI portfolio.
Outcomes
Six-step eligibility check across Karnataka and central schemes
English and Kannada throughout, toggled in one tap
WhatsApp assistance for citizens who won't use a web form
Application tracking across preparation, verification, assignment and submission
DPDP Act 2023 privacy policy, terms and data-deletion route
Deal flow intelligence for financial institutions.
Debt VC firms, banks and NBFCs drown in deal flow. Hundreds of opportunities arrive every month through email, warm intros and broker networks, and most never get properly evaluated.
AURA structures, scores and tracks every opportunity from first touch to exit — ingesting memos, decks, financials and market data into one deal profile, with an underwriting workspace for versioned documents, checklists and approvals.
Built as white-label SaaS for institutions that want modern deal intelligence without building it themselves. Multi-tenant, API-first, designed against enterprise security and compliance requirements.
Outcomes
Deal scoring from unstructured input — decks, memos, financials
Unified pipeline from origination to portfolio management
Collaborative underwriting workspace with approval workflows
Portfolio monitoring with early warning signals
White-label, multi-tenant architecture
Enterprise SaaS · in stealth
gallery
Other things I've built
Weekend builds, internal tools and landing pages. Some shipped, some scratched an itch.
timeline
The path here
Twenty-two years across enterprise and startups — technology, product, execution.
2022 — Present
Founder journey · Tipplr → Nexus AI / CIQ AI Solutions
Co-founded Tipplr in January 2022, scaled it into one of India's leading ONDC food networks and led it through a successful acquisition in 2026. Now building AI-first products through Nexus AI / CIQ AI Solutions — AskGogo, CreditIQ, Rasoi Capital, FolioKey and AURA.
2018 — 2021
Infrastructure Head · JCPenney Services India
Owned global infrastructure design and deployment for the retailer's India technology centre — network, data centre and unified communications across sites. Now Catalyst Brands India.
2015 — 2018
Lead Architect, Voice · Capita IT Enterprise Services
Lead UC architect on Cisco Collaboration engagements — architecture, statements of work and delivery for global clients across India and UK centres.
2011 — 2015
AVP, Voice Engineering · JPMorgan Chase
Carrier-grade voice architecture for global trading floors — HLD/LLD, capacity planning, SLA and monitoring design across cloud and on-premise.
2010 — 2011
Sr. Technology Analyst · Goldman Sachs
Unified communications infrastructure for the firm — Cisco IPCC, CUCM, CUC and CUCCX platforms.
2009 — 2010
Project Lead · Target
Project planning, systems design and deployment, cutover and operations — plus service improvement and team performance management.
2004 — 2008
Sr. Voice Consultant · Accenture
End-to-end voice infrastructure design, implementation and maintenance for contact centre operations.
2003 — 2005
Sr. Executive Engineer · Siemens
Installation and maintenance of HiPath EPABX, contact centre, voicemail and video conferencing systems. Where it started.
process
How I work
The systems behind the output. The case studies show what got built — this is how.
The shape of it
I don't build products as a series of features. I build them as loops — systems that pick up a problem, run it through a defined process, check their own output, and leave me with a decision instead of a draft. Every product on this site follows the same five pieces, arranged differently: problem identification, validation, build, measure, iterate.
Mock before docs
For anything with a UI, I want a clickable mockup approved before a single spec is written. A document can describe a flow; it cannot show whether the thing in your head is actually the thing on screen.
Six docs, then phase by phase
PRD, TRD, app flow, UI/UX brief, backend schema, implementation plan — then build in small phases. I don't ask an AI or a team to swallow an entire application in one shot.
The builder never reviews its own work
One AI builds; a different one reviews the plan and the diff. Fresh context catches assumptions the builder has already normalised.
Accepted is not delivered
An API returning 200 proves acceptance, not the outcome. Every meaningful change gets verified in the actual production path before I call it done.
The loop closes in public
LinkedIn for reach, this site for depth. Every product that performs gets written up here in full — the honest numbers, the failures, the pivots. The loop only closes when the learning is shared.
What I look for in problems
High frequency of pain — daily beats monthly
Existing spend — people already paying for something worse
A structural tailwind — ONDC, AI adoption, digital India
A defensible data flywheel — every user makes it better for the next
Clear unit economics — a visible path to profitability
Stack & tools
Next.js, React, TypeScript on the front end
Supabase and PostgreSQL for data
Claude Code and OpenAI for AI features and build workflows
Twilio + WhatsApp Business for messaging surfaces
Razorpay for payments
pgvector for retrieval and memory
Vercel for deployment
advisory
Build with me
I spend most of my time building Nexus AI / CIQ AI Solutions. I selectively work with founders and teams where my operating experience can materially change the outcome.
AI product development
Prototype to production. What to build versus buy, where a model genuinely beats a rule, and how to ship an AI feature people use on day thirty rather than day one. I've taken five products through this, so I know which parts are hard and which only look hard.
Fintech product strategy
Credit, lending and investment products under Indian regulation. What DPDP compliance actually costs to build, how to price honestly when every competitor is paid to rank, and where fintech trust is won — usually in the boring parts nobody demos.
Distribution: WhatsApp, ONDC, vernacular
The channels most founders treat as an afterthought and then bolt on badly. What worked scaling Tipplr on ONDC from 300 to 2,000 daily orders, what building AskGogo taught me about WhatsApp as a product surface, and why vernacular is a distribution decision rather than a localisation task.
Operating AI-native
Running a portfolio of products with AI doing most of the building. The systems that make it work, the failure modes that cost me weeks, and the places a human still has to sit in the loop. Honest about what breaks, not just what ships.
Building something where this experience is useful? Say hello.
writing
Essays & notes
Practical notes on AI-native building, fintech product strategy, and what shipping actually teaches you.
about
A bit more
I started as a technology analyst. Then I found product. Then I found entrepreneurship. The arc isn't linear, but the through-line is: I like making things that move outcomes — and AI is the most leveraged way to do that right now.
I co-founded Tipplr in January 2022, built it into one of India's leading ONDC-powered food commerce platforms and led it through a successful acquisition in 2026. Before that I led voice architecture at Capita IT, was AVP of Voice Engineering at JPMorgan Chase, a Senior Technology Analyst at Goldman Sachs, and a Project Lead at Target. Twenty-two years across enterprise and startup, technology and product, execution and strategy.
Now I build through Nexus AI / CIQ AI Solutions — a portfolio of AI-native products including AskGogo, CreditIQ, Rasoi Capital, FolioKey and AURA. The operating model is deliberately small, high-leverage and AI-first: find a meaningful problem, understand the system behind it, ship quickly, and keep tightening the loop with real usage.
I'm looking for founders, investors, operators and builders who care about the same things. If that's you, say hello.
contact
Say hello
Hiring, advising, investing, or just comparing notes on building with AI — all welcome.
Opens your mail app with the message ready to send.
AI-native builder with 22+ years shipping products end to end. I don't hand off specs — I close the gap between idea and shipped product. I've gone from technology analyst to production code, owned growth funnels, and taken 0→1 features to market. Since 2022 I have built and operated my own products full-time.
I build through Nexus AI / CIQ AI Solutions, using Claude Code and an AI-native workflow to move from problem definition to shipped product and adoption.
Live work
AskGogo
askgogo.in — Personal AI agent across WhatsApp and web — memory, approvals and background runs
CreditIQ
creditiq.app — Honest credit card intelligence, real math not marketing
Tipplr
ONDC food network — built, scaled and successfully acquired
