// Player profile · New Delhi, India

Vrishank Gupta_

Building AI products that ship — engineer's depth, product manager's judgment.

Class: AI Product ManagerLevel 28Guild: Hero ElectronixAlum: ISB Hyderabad Alum: NSIT CoE

OriginSoftware Engineer→AI Product Manager· hands-on with the stack

Engineer-turned-product-manager: I still read the code and the architecture, then turn ambiguous problems into AI products customers love and businesses value — from LLM chatbot platforms and RAG systems to workflow automation used by millions.

LLMsRAGAI AgentsConversational AIGenAI AnalyticsMCP ArchitectureNeural Networks

Explore the quest log Get in touch Résumé PDF

Vrishank Gupta

Status · Shipping

AI Product Manager · New Delhi, India

AI Vision92

Product Craft88

Code Sense84

// Impact points earned

Tickets self-served by AI chatbot

Chatbot flows shipped across 40+ SKUs

−0 hrs

Resolution time after CRM + telemetry fusion

−0%

Field visits via remote diagnostics

₹0L

Upsells from AI routing engine

−0%

Inbound support calls

// Chapter I

Quest Log

Seven campaigns across AI product management and engineering — from national-scale government tech to LLM platforms serving millions.

Hero Electronix

● ActiveMain Quest

Product Manager

Jul 2025 — Present

Gurugram, India

+12,000 XP

Building the AI self-service brain behind Hero's smart home devices — chatbots, telemetry, and LLM analytics in one loop.

▸AI chatbot became the primary support channel — 58% of monthly tickets deflected, inbound calls down 26%
▸Shipped 320+ chatbot flows across 40+ SKUs for self-troubleshooting and feature guidance
▸Fused CRM + real-time device telemetry + ticketing: TAT down 23 hrs, field visits down 48%, device replacements down 31%
▸Launched an LLM-powered analytics interface over MySQL — leadership queries ops metrics without BI dependency
▸Ran POCs with 8 conversational-AI vendors; drove 4 hands-on LLM enablement sessions with engineering
▸Shipped AI alerts on cameras & video doorbells — person detection, intruder alerts, motion intelligence
▸Built an in-house Amazon review scraping + AI analytics pipeline, consumed directly by CXOs

Impact:58% self-serve−26% calls−23 hrs TAT−48% field visits−31% replacements

Darwinbox

✓ CompleteCampaign

Product Manager

Apr 2024 — Jul 2025

Hyderabad, India

+11,000 XP

Two campaigns: the Bots & Helpdesk platform, and Employee Experience (Surveys & Engagement) for a global HR SaaS.

▸Unified chatbot backend across web, mobile, MS Teams & Slack — one brain, every channel
▸AI rules engine auto-classifies and routes tickets; first-response SLAs visibly improved
▸Blended LLM response engine merging standard intents, FAQs and policy search into one intelligent layer
▸AI FAQ generator trained on historical tickets — policy documentation goes from weeks to hours
▸Intelligent routing engine drove 3 upsells worth ₹23 Lakhs
▸Leading a company-wide POC on MCP-server architecture for the bot's real-time client stack
▸8 enterprise demos, 7 pilots, and a webinar with 44+ clients → 17 qualified leads
▸Unified survey UX converted 12 paid clients; Moments-based surveys closed 3 more post-demo
▸Attrition-Risk & Manager-Performance dashboards; confidentiality overhaul lifted adoption 23%

Impact:₹23L upsells17 qualified leads12 clients converted+23% adoption

Adobe

✓ CompleteCampaign

Member of Technical Staff 2

Jan 2022 — Apr 2023

Noida, India

+7,500 XP

Document Cloud division — backend APIs powering PDF operations for developers worldwide.

▸Backend APIs for PDF operations in Adobe Document Cloud
▸Built the Power Automate connector for PDF Accessibility — Spot Award for the Microsoft collaboration
▸Built a backend microservice from scratch powering try-it-out UIs for Adobe's API catalog
▸Mentored fellow developers across multiple domains

Impact:Spot AwardMicrosoft collab

Flipkart

✓ CompleteSide Quest

Software Development Engineer

Jun 2021 — Jan 2022

Bengaluru, India

+3,000 XP

Supply-chain systems in Ekart's e2e-planner team — through Big Billion Days scale.

▸Systems managing shipment flow and planning across Ekart's supply chain
▸Multi-Pickup-for-Sellers → +3% D2 speed on NFBF shipments
▸Java, Spring Boot, MySQL, Grafana — distributed services built for peak-sale events

Impact:+3% D2 speedBig Billion Days

MakeMyTrip

✓ CompleteSide Quest

Software Engineer

Jul 2020 — Jun 2021

Gurugram, India

+3,500 XP

Hotels platform backend — search, scale, and a new language for the Gulf.

▸Hotels & Locus backend microservices — Spring, Node.js, MongoDB
▸Migrated architecture to support Arabic for GCC countries on MakeMyTrip
▸ElasticSearch pipeline for Hotels Autosuggest
▸Delivered static content: amenities, locations, videos & imagery

Impact:GCC launchAutosuggest pipeline

Publicis Sapient

✓ CompleteSide Quest

Product Engineering Intern

May 2019 — Jul 2019

Gurgaon, India

+800 XP

One dashboard to watch the whole delivery pipeline.

▸Web dashboard aggregating JIRA, Bitbucket, Jenkins & Zephyr into a near-real-time view of the dev pipeline

Impact:First shipped dashboard

NITI Aayog

✓ CompleteSide Quest

Software Engineering Intern

Jun 2017 — Jul 2017

New Delhi, India

+500 XP

Government-tech origin story at NIC under the Atal Innovations Mission.

▸Built the official submission website for Atal Tinkering Labs entries from schools across India
▸Authored a project report on border-defence tech — LASER fencing, GPR, thermal sensing

Impact:National rollout

// Case files

Case Studies

Deep dives into the most relevant AI product work — the problem, the approach, and what actually moved.

AI self-service as the primary support channel

Hero Electronix · Product Manager

The problem

Inbound support calls were drowning the contact centre for smart home devices. Customer context was fragmented across call-centre, field-service and ticketing systems, so every resolution cycle dragged and field visits were the default fix.

My approach

01Designed and shipped 320+ chatbot flows across 40+ SKUs, covering self-troubleshooting and feature guidance so customers never needed to call for common issues
02Fused CRM with real-time device telemetry and the ticketing system — agents see device state and error history before the conversation starts
03Unified schemas and lifecycle states across call-centre and field-service platforms to end fragmented CX visibility
04Ran POCs with 8 conversational-AI vendors to pick the stack, then drove 4 hands-on LLM enablement sessions with engineering

Measurable outcomes

58% of monthly tickets self-servedInbound calls −26%Resolution time −23 hrsField visits −48%Device replacements −31%

Conversational AILLMsDevice TelemetryCRM Integration

One conversational AI platform for every channel

Darwinbox · Product Manager

The problem

A global HR SaaS was running fragmented bots across web, mobile, MS Teams and Slack. First-response SLAs suffered, policy documentation lagged, and leadership couldn't see what clients actually needed from support automation.

My approach

01Spearheaded a unified chatbot backend so one brain serves every channel with consistent UX
02Delivered a blended LLM response engine merging standard intents, FAQs and policy search into a single intelligent layer
03Built an AI rules engine for automated ticket classification and routing, and an AI FAQ generator trained on historical tickets
04Led a company-wide POC of MCP-server architecture for the bot's real-time client communication stack
05Supported GTM directly: 8 enterprise demos, 7 pilots and a webinar with 44+ clients

Measurable outcomes

₹23 Lakhs upsells from AI routing17 qualified leadsFirst-response SLAs improvedPolicy docs from weeks to hours

LLM OrchestrationRAGAI AgentsMCP Architecture

Ask the database: LLM analytics for leadership

Hero Electronix · Product Manager

The problem

Leadership needed operational answers — ticket trends, field-service efficiency, device failure patterns — but every question meant a BI ticket and days of waiting, so decisions ran on stale dashboards.

My approach

01Designed an LLM-powered natural-language analytics interface over MySQL, mapping operational metrics into a semantic layer the model could reason about
02Guardrailed query generation so leadership gets trustworthy numbers without writing SQL or filing BI requests
03Paired it with an in-house Amazon review scraping + AI analytics pipeline, giving CXOs voice-of-customer trends alongside ops metrics

Measurable outcomes

Zero BI dependency for ops queriesNear-real-time leadership decisionsCXO-grade review analytics

LLM AnalyticsMySQLSemantic LayerAI Pipelines

// Chapter II

Skill Tree

Three branches leveled across the journey — each node forged somewhere in the quest log above.

AI & Machine Learning

GenAI & LLMs
Conversational AI
RAG Pipelines
AI Agents
Neural Networks
Prompt Engineering
Model & Vendor Evaluation
MCP Architecture

Product

Roadmapping
0 → 1 Building
Stakeholder Management
Experimentation & Data
Customer Research
Metrics & Analytics
GTM, Demos & Sales

Engineering

Java & Spring Boot
Microservices
MySQL / SQL
Node.js
MongoDB
ElasticSearch
Distributed Systems
Git & Grafana

// Chapter III

Training Grounds

Two institutions, two layers of the same build — the engineering foundation, then the business layer on top.

ISBThe business layer

2023 — 2024

Indian School of Business

PGP (MBA) · Marketing & Strategy double majors

Where the engineer learned to price, position and sell — turning customer problems into revenue and roadmaps into GTM plans.

Pricing & GTM strategy
Marketing & Strategy double majors
Class of 2024

NSITThe technical foundation

2016 — 2020

Netaji Subhas Institute of Technology

B.E. Computer Engineering

Four years of computer engineering built the depth I still use daily — I review architecture diagrams before I sign off on a roadmap.

Computer Engineering core
Distributed systems foundations
Smart India Hackathon finalist

The combination is the point — engineer's depth, product manager's judgment. I can review the architecture diagram and the pricing model in the same meeting.

// Side quest

Interrogate the build

A tiny local model, pre-loaded with everything on this page. Pick a query — no network calls, no hallucinations.

vrishank@portfolio:~/ai-terminal

❯ awaiting query… select one below█

Explore the work

Recruiter mode

// Chapter IV

Trophy Room

Achievements unlocked and licenses earned along the way.

Smart India Hackathon

National Finalist · 2019

Epic

WRLDS Hackathon

Singapore

Epic

Ace Alliance

Flipkart

Rare

Spot Award

Adobe · Microsoft collaboration

Rare

Knowledge scrolls

Certifications

✦Neural Networks and Deep Learning
✦Improving Deep Neural Networks: Hyperparameter Tuning, Regularization & Optimization
✦Brand and Product Management
✦Managing the Company of the Future
✦Marketing in a Digital World

// Final chapter

Start the conversation

Looking for someone who pairs deep technical understanding with strong product thinking? Send a party invite.

Party invite open

Zero-to-one AI products, LLM platforms, enterprise SaaS — or just a good conversation about where conversational AI goes next.

Emailhello@vrishankgupta.in Phone+91 91366 66425 LinkedInlinkedin.com/in/vrishank-gupta GitHubgithub.com/Vrishank-Gupta

BaseSarita Vihar, New Delhi