E-Commerce Growth  ·  Marketplace & Seller  ·  AI Agents

Gagandeep Goyal

I turn traffic and campaigns into revenue — for e-commerce brands.

E-commerce operator with a marketplace-growth and performance-marketing edge and 10+ years across commerce, online and offline. Founder of IndiaOffers.in — now building AI agents that put a decade of marketplace experience into software.

About

Commerce, from both sides of the counter.

I grew up inside my family’s retail & wholesale business in Punjab — learning sourcing, pricing, margins, and what actually makes a customer buy, long before I ever touched a marketplace dashboard.

Since 2011 I’ve lived in e-commerce: as a founder, a marketplace & affiliate operator, and a power-user across virtually every major Indian marketplace. That combination gives me a rare, ground-level read on both sides of the transaction — buyer behavior and pain points, and the seller economics behind them: margins, inventory, RTO, catalog, and trust.

Today I run performance channels end to end — onboarding networks and brands, launching campaigns, screening out fraudulent traffic — and I ship my own products. After Harvard’s CS50 that meant web apps; now it means AI agents built on Google’s Agent Development Kit that encode everything the decade taught me about how marketplaces really work.

Based in
Punjab, India · Remote & Global
Open to
Marketplace & Seller Growth · Affiliate & Partnerships · Performance Marketing · E-Com Ops
Education
B.Tech, Industrial Engineering
NIT Jalandhar · CGPA 8.07
Also fluent in
Building AI agents & web products — Google ADK · Gemini
0
Across e-commerce, marketplaces & retail
0
AI agents built on Google’s ADK — Vertex AI-ready
Both
Sides of commerce — buyer & seller, online & offline
0
Web & AI products shipped solo, end to end

AI Agents

Ten years of marketplace experience, compiled into software.

Three production-grade agents built on Google’s Agent Development Kit (ADK) with Gemini, each deployable to Vertex AI Agent Engine. Their playbooks weren’t prompted from thin air — they encode a decade of doing the job: running stores, fighting RTO, chasing refunds, and reading buyers.

Marketplace Seller Agent

AI team for marketplace sellers

A Store Manager agent plus four specialist sub-agents that run a seller’s store across Amazon.in, Flipkart, Meesho & Myntra.

  • Coordinator + 4 specialists — pricing, inventory, listings, returns — each with its own tools and playbook.
  • Talks net profit, not gross: full fee model with commission, closing fees, shipping, GST and a return-cost reserve.
  • Account-health guardian — tracks late-shipment, cancellation and defect rates against suspension thresholds, per marketplace.
  • Opens every day with a store health digest; prices per marketplace, never below the margin floor.
Google ADKGeminiMulti-agentVertex AI-ready
View on GitHub

Marketplace Support Agent

Customer support, marketplace-style

An Amazon/Flipkart-style customer-support agent with dedicated order and returns specialists and a clean path to a human.

  • Coordinator routes each query to an order agent (tracking, order details) or a returns agent (eligibility, refunds, cancellations).
  • Policy-aware — decisions checked against store policy tools, with human escalation built in.
  • Swappable data layer designed for real marketplace APIs — Amazon SP-API, Flipkart Seller API, Zendesk.
  • Ships with a deploy script for Vertex AI Agent Engine.
Google ADKGeminiSub-agentsVertex AI-ready
View on GitHub

Complaint Write Agent

Consumer-grievance advocate

Interviews a customer one question at a time, gathers evidence, and drafts natural, human-sounding complaint emails that get resolutions.

  • Attempt-based tone ladder — friendly on first contact, firm and fully detailed by the fourth.
  • Evidence-first — asks for screenshots before drafting and references them in the email.
  • Zero-placeholder, ready-to-send drafts with every detail collected up front.
  • Trainable memory — ships with a 25-rule playbook and learns new durable preferences from plain chat.
Google ADKGeminiPersistent memoryVertex AI-ready
View on GitHub

Built AI-first, by design. These agents were developed hands-on with AI pair-programming — a workflow I trained myself in — while the architecture, the marketplace playbooks and every product decision are my own. All three codebases are public on GitHub, open to review.

Experience

A decade of operating, building, and selling.

2015 — Present · Remote
Founder & Operator — IndiaOffers.in
  • Built and run a cashback & deals platform monetized through affiliate partnerships with networks and direct brands — Flipkart, Amazon, CJ, Admitad, Cuelinks, vCommission, INRDeals.
  • Drove millions of rupees in tracked sales for partner networks; own the full lifecycle — onboarding, campaigns, attribution, payouts, and fraud-traffic screening.
  • 2024–25: rebuilt the platform ground-up with AI-assisted development. 2026: building AI agents on Google’s ADK.
Long-standing · Punjab, India
Retail & Wholesale Operations — Family Business
  • Hands-on with sourcing, pricing, margins, inventory, and daily face-to-face customer service.
  • First-hand grasp of offline buyer behavior and small-seller economics — the same pain points that shape how people buy and sellers survive online.
Dec 2015 — Jul 2016 · Mumbai
Program Associate — Avanti Learning Centres
  • Taught JEE Main/Advanced students with peer-learning pedagogy (developed with Prof. Eric Mazur, Harvard); owned batch performance and curriculum feedback.
2015 · Odisha
Management Trainee — Bhushan Steel Ltd.
  • Campus-placement rotational training across departments; hands-on floor experience in the Hot Strip Mill.
May — Jul 2014 · Gurgaon
Summer Intern, Operational Excellence — Hindustan Coca-Cola Beverages
  • Applied Linear Programming to a supply-chain problem; completed Level-1 Lean & Six Sigma; contributed to a yield-improvement project.

How I operate

Hard problems don’t survive sustained attention — so that’s what I give them.

First principles, always

I take a problem apart before choosing a path. The goal is the right solution — not the first one that works.

Depth over hours

When something matters, I stay with it long after the workday ends. My best solutions have usually arrived after everyone else stopped thinking about the problem.

Thinking ends in shipping

Analysis is only worth what it produces: a campaign, a feature, an agent — live, and measured against real numbers.

Capabilities

What I bring to the table.

E-Commerce & Marketplace

Marketplace & seller ops Seller onboarding & growth Catalog / SKU RTO reduction Last-mile & fulfillment Customer experience

Affiliate & Partnerships

Network & brand onboarding Campaign management (CPS / CPL / CPI / CPC) Publisher relations Cashback & coupon models Payout reconciliation

Performance Marketing

Conversion optimization Attribution & tracking Fraud / fake-traffic detection Mystery shopping Competitor benchmarking

AI, Analysis & Building

AI agents (Google ADK · Gemini) Multi-agent systems Vertex AI Agent Engine Python Operations Research Lean & Six Sigma HTML / CSS / JS

Networks & platforms

Admitad vCommission Cuelinks INRDeals CJ (Commission Junction) Flipkart Affiliate Amazon Associates

Contact

Let’s build something that sells.

Looking for a marketplace-growth, affiliate, or performance operator who knows commerce inside out — and builds AI agents too? Let’s talk.