Founder
Sunil Kumar
Founder, TwoDots · AI Implementation for E-Commerce
I spent 15 years building AI inside Kohl's and Sears — recommendation engines, demand forecasting, inventory optimisation — before coming back to India and realising the same playbook could work for a $5M Shopify brand. That's what TwoDots does.
Years in retail AI
15+
Retailers worked with
Kohl's, Sears
Sellers on HappySellers
250+
Registered HappySellers businesses
6,000+
WORLDEF
Exhibitor
Based
Pune, India
Background
I started my career as a software engineer in India, then moved into the US retail technology space — first at Sears, then at Kohl's Department Stores in Milpitas, California, where I spent a decade building AI systems at enterprise scale. I led the recommendation engine that served millions of shoppers across Kohl's website and stores, survived multiple Black Friday deployments, and learned something most AI consultants never learn: what it takes to ship AI that actually works under production load.
In 2018, I came back to India. I saw the e-commerce market exploding — Shopify, Flipkart, Amazon, D2C brands — and I saw the same operational problems that big retailers had solved with AI, playing out at 1/100th the budget and 100x the manual effort. Sellers were running on spreadsheets, losing 15–20% of potential revenue to stockouts, drowning in manual reconciliation, and watching returns eat margin they'd never get back.
We built HappySellers to solve this. Today it handles order management, inventory tracking, and payment reconciliation for 250+ active sellers and over 6,000 registered businesses. It's not a side project — it's the live laboratory that makes TwoDots' implementation advice credible. Every AI pattern we recommend has been tested on real GMV first.
TwoDots was the natural next step: a firm that takes the enterprise AI playbook and compresses it into 4–12 week implementations for $1M–$20M online sellers. No decks. No pilots. Working AI in production — or we don't charge for the milestone.
Three things I'm known for saying
Story 1
"The recommendation engine that paid for itself by lunchtime."
At Kohl's, we launched a new recommendation model on a Tuesday morning. By lunch, it had covered the entire project cost in incremental revenue. The same math works for a $5M brand.
Story 2
"250 sellers, one dashboard, zero hype."
HappySellers is a live lab, not a portfolio. When I recommend an AI approach, I'm not guessing — I've seen it run on real SKUs for real sellers in real seasons.
Story 3
"The $5M brand that got a Kohl's-grade playbook in 6 weeks."
Demand forecasting, reorder optimisation, returns prediction — the same tools Kohl's paid millions for, shipped into a D2C brand in six weeks at a fraction of the cost.
Speaking topics
Available for conferences, podcasts, webinars, and panel discussions. Has exhibited at WORLDEF.
Book a speaking slot →AI Implementation for E-Commerce SMBs
The practical playbook — what to implement first, what data you actually need, and how to measure ROI in the first 30 days.
From War Room to Shopify Admin: Lessons from Kohl's Black Friday AI
How enterprise-grade AI at Black Friday scale translates to a Shopify brand. The patterns that work at $10B also work at $5M.
Why AI Projects Fail — And How to Make Them Ship
70% of AI pilots never go live. Here's the anatomy of a failed AI project and the implementation discipline that changes the outcome.
Demand Forecasting for Online Sellers: A Practical Guide
How to reduce stockouts and overstock simultaneously using models that work on the data you already have.
The $5M AI Playbook — Kohl's-Grade Results on an SMB Budget
A concrete implementation roadmap for e-commerce brands doing $1M–$20M. Prioritised by ROI, sequenced by data readiness.
HappySellers: Building a Live AI Lab for 250+ Online Sellers
How a platform built for Indian sellers became the world's best laboratory for testing AI in e-commerce operations.
The honest version
"I help $1M–$20M online sellers ship AI that actually works.
I built the recommendation engine at Kohl's; now I do that for Shopify brands."
— the 10-second version, unchanged since day one