Hi, I’m
Swathi
Chandrasekaran.
Product Manager at Unilever. I build AI products that run across 20+ markets.
Three years at NatWest taught me how products get built. Two years at XLRI taught me why they should. Now I do both at once.
From code to product.
Building products that transform customer operations for one of the world's largest consumer goods companies.
One of a small cohort chosen for Unilever's flagship leadership programme, rotating end-to-end across the FMCG value chain – factory floor to the consumer's shelf.
Two years at XLRI – one of India's most rigorous business schools – because I wanted to own the problem, not just build the solution.
Three years transforming the personal banking journey for customers at scale – the foundation everything else is built on.
Where Hello World was just the beginning. Four years at one of India's oldest engineering colleges turning 0s and 1s into actual software.
From writing code to owning the product.
Product Manager
CurrentUnilever · May 2025 – Present · Bengaluru, India
- Transforming Unilever's global order capture with agentic AI — automating ~300K annual orders across 20+ markets, owning the roadmap and E2E rollout.
- Delivered MVP cutting onboarding time by 50% and releasing 10+ FTE; learning agent hit 98% extraction accuracy.
- Revamped agentic infrastructure to cut cost-per-order by 40%; automated 40% of test-case execution via SonarQube to accelerate release cycles.
Unilever Future Leaders Programme (UFLP'24)
Unilever · May 2024 – May 2025 · India
- Rotated across 8+ high-impact projects spanning 5+ FMCG domains — AI assistants, last-mile ops, consumer journeys, and more.
- Shipped Innoflex GPT from 0-to-1: prompt engineering and testing drove +30% weekly active users.
- Cut process time by 40% by digitising Lakme hardware capex allocation, enabling end-to-end asset lifecycle tracking.
- Improved last-mile efficiency by 20% through field ops mapping and 15+ sales beats across markets.
Software Engineer
NatWest Group (formerly RBS) · Jul 2019 – Jun 2022 · Chennai, India
- Drove ₹2M in savings — built an end-to-end mortgage data transfer system for personal banking customers.
- Delivered ₹1.5M savings — redesigned and streamlined the personal banking account-opening journey.
- Cut deployment time 8× (4 hours → 30 minutes) by building a fully automated CI/CD pipeline from scratch.
Two institutions. One career.
XLRI Jamshedpur
PGDM (BM)
Business Management
College of Engineering, Guindy
B.E. Computer Science
Anna University
A few things the CV skips.
I left home when I was 17 – first time away, first time figuring things out alone.
Turns out leaving your comfort zone is a habit that sticks. I’ve been jumping in head-first ever since – through engineering, an MBA, and now product at Unilever. I think in systems and I’m drawn to the human side of every decision. When I’m not solving product problems, I’m growing a garden that demands patience, painting things I can’t explain, or baking cakes for anyone nearby.
📍 Bengaluru, India
I don’t just spec products – I build them.
A few things I’ve shipped on nights and weekends – mostly AI agents and integrations that scratch a real itch.
Web Search Agent with Voice & Memory
I wanted a personal assistant that actually remembers me. I gave it simple tools to start, then added voice because I'm too lazy to type — the eventual goal is something that just knows me.
An agent that evolved from a single Google Search tool into a full voice assistant — takes a call via Twilio, recalls who you are from Mem0, reads full web pages, scores job listings against your resume, and emails results. Same tool layer, text or voice.
AI Job Finder
By the time I found a job on LinkedIn, it already had 100+ applicants. Hours of scrolling, low odds. I built something that brings the right roles to me daily — before the crowd gets there.
Wake up to your top 10 matching jobs in your inbox — the agent runs overnight, scours job boards, and delivers only the roles that fit.
Swiggy MCP Server
I wanted to order food just by talking. Built this as a conversational Swiggy interface — and I'm working on plugging it into the personal assistant so it can handle food on command.
A Model Context Protocol server that plugs Swiggy into any AI assistant — browse, search and order food entirely through natural language.
Telegram AI Bot
I wanted Claude accessible where I already spend time. Same model, different channel — Telegram felt like the most natural place to have a conversation.
An AI assistant that lives inside Telegram — answering questions and running tasks right where the conversation already happens.