Product Manager  ·  London, UK  ·  9+ years

Manual to automated.
Legacy to platform.
AI from the infrastructure up.

Product Manager with 9+ years building the systems underneath the surface — platforms, data infrastructure, and AI at scale. The foundations that make everything else possible.

3–6 mo → 1 wk AI model deployment time cut at Sportradar
23 models Migrated to cloud AI hosting platform in 7 months
£21M+ Revenue unlocked across products at THG
80% Engineering overhead eliminated via automation
AI in my work
Platform builder: Led cloud infrastructure hosting 23 analytical AI models at Sportradar — version control and governance introduced from scratch. AI practitioner: Used AI tooling to ship a live product prototype before a single line of production code was written — cutting discovery time and accelerating stakeholder alignment.
AI platform products AI-assisted prototyping Platform engineering Legacy modernisation API products eCommerce & payments Data platforms Self-service systems Cambridge MBA

About

Most people talk about solving problems. I'm more interested in the uncomfortable silence before anyone knows what the problem actually is.

I started my career in sales — which turned out to be the best possible training ground for product management. Before you can sell anything, you have to understand the person in front of you. Their real frustration, not the one they've articulated. Their real goal, not the one written in the brief. That instinct never left me.

The move into product felt less like a career change and more like a natural evolution. Sales taught me how to listen without an agenda. Product gave me a way to act on what I heard — at scale, in systems, with engineering teams who needed clarity and stakeholders who needed conviction.

Today I build platform products at Sportradar — infrastructure-level work that rarely gets a standing ovation but keeps everything else running. I've migrated AI models to the cloud, replaced legacy systems with self-serve platforms, and used AI tooling to ship prototypes faster than anyone expected.

Outside work, I lift. I run. I'm a family man. The discipline of training follows you everywhere — you learn to trust the process when results aren't visible yet, and you get comfortable doing unglamorous work well.

Origin story
Sales → Product
Business development taught me to listen without an agenda. Product gave me a way to act on it at scale.
How I think
Problem first, always
Before roadmaps or solutions — understand what is actually broken, for whom, and why it matters right now.
AI perspective
Builder & practitioner
I've built AI infrastructure at scale and use AI day-to-day as a product tool. Both sides of the coin.
Outside work
Lifting. Running. Family.
Training teaches you that meaningful progress is invisible for a long time. That applies directly to platform work.
Education
MBA, Cambridge
Judge Business School · 2022–2023. B.Tech Chemical Engineering, VIT · 2012–2016.

Work

The work that shaped how I think

Two deep dives into platform products built at Sportradar — both from the same conviction that the legacy way wasn't good enough. Below, snapshot cards covering the broader body of work across THG, PQE, and BYJU'S.

AI Platform · Sportradar · Deep dive

AI Model Hosting Platform — from local machines to a shared cloud brain

Reimagining how analytical AI models are deployed and consumed — from isolated stateless models on individual machines to a centralised, governed cloud platform accessible via API and real-time streams.

The problem
  • AI models lived on individual machines with no centralisation
  • No versioning — model quality entirely uncontrolled
  • Deploying one model took 3–6 months of manual effort
  • Every team managed their own models independently
What I did
  • Led end-to-end reimagination of the model hosting architecture
  • Drove migration from stateless local to stateful cloud infrastructure
  • Established model governance and versioning from scratch
  • Enabled self-serve consumption via APIs and real-time streaming
  • Coordinated across engineering, data science, and operations
The outcome
  • Deployment cut from 3–6 months to 1 week
  • 23 of 34 models migrated to cloud in 7 months
  • First-ever version control and quality governance
  • Multiple teams consuming shared models via API and stream
1 weekDeployment time (was 3–6 months)
23/34Models migrated in 7 months
FirstModel governance established
API + streamConsumption methods enabled
Platform · Sportradar · Deep dive

Markets Catalogue — making the source of truth actually usable

A centralised data catalogue replacing a fragmented legacy system, with a self-service management layer and a human-facing frontend — prototyped using AI before a single line of production code was written.

The problem
  • No single source of truth for market data
  • Market launches took days or months with heavy engineering dependency
  • No self-serve capability — everything went through engineers
  • API-only architecture inaccessible to business users
What I did
  • Defined vision and led delivery of the centralised catalogue
  • Designed multi-consumer API service layer for downstream systems
  • Ran discovery, built MoSCoW-prioritised frontend requirements
  • Used AI to deploy a live Vercel prototype before committing to build
  • Implemented auto-scaling for peak resilience
The outcome
  • Market launch cut from days/months to minutes
  • Business users fully self-serve — zero engineering dependency
  • Multiple downstream consumers via purpose-built endpoints
  • Stakeholders validated live prototype before build began
MinutesMarket launch (was days/months)
ZeroEngineering dependency for routine tasks
LiveAI-built prototype pre-build
MultiDownstream consumers enabled
THG · Checkout & payments
Intent funnel platform
Owned basket, checkout, offers, payments, and fraud prevention. Launched BNPL, retail media ads, and gift cards as major revenue initiatives.
BNPL conversion uplift+3.6%
Gift card revenue£5M+
Ad revenue increase+7%
THG · Internal tooling
CRM & offers tooling
Built internal CRM from scratch, rolled to 12 brands. Redesigned offers UI across 23 offer types and 120+ brands — cutting costs and engineering overhead significantly.
CRM cost saving£0.31M
Offers tooling saving£0.4M/yr
Engineering overhead cut-23%
PQE · SaaS platform
ServiceNow automation
Automated database build for pharma compliance SaaS. Streamlined change management tooling across 12 new features and drove data-led UX uplift.
Engineering overhead cut-80%
Incident resolution time-33%
+42%Conversion uplift

Experience

Sportradar
Jan 2025 – Present · London
Product Manager — Platform
Building the next-generation platform replacing a legacy system — unifying siloed components into a modular, scalable architecture. Leading AI Model Hosting and Markets Catalogue.
The Hut Group (THG)
Sept 2023 – Dec 2024 · London
Product Manager — eCommerce Platform
Led the intent funnel for a £2.2B global eCommerce business — basket, checkout, payments, offers, and fraud. Launched BNPL, retail media, and gift card products. Built and shipped the internal CRM.
Cambridge MBA
2022 – 2023
Judge Business School, University of Cambridge
Full-time MBA — strategy, leadership, and general management across a global cohort.
PQE Group
Nov 2018 – Jul 2022 · India & Italy
Senior Product Consultant
Delivered pharma compliance SaaS products. Automated core platform processes, improved client tooling, and drove data-led UX improvements across the web platform.
BYJU'S
Jun 2016 – Nov 2018 · India
Senior Product & Business Development Manager
Started in sales — the foundation of everything. Moved into product, improving checkout pay-complete rate by 14% and building the instincts that define how I work today.
AI & platform
AI platform products AI-assisted prototyping Platform engineering Legacy modernisation API product management Data infrastructure Self-service systems
Product craft
Product vision & strategy Roadmap & prioritisation Discovery & user research OKR definition Go-to-market strategy Agile & Scrum Stakeholder alignment A/B testing SQL · JIRA · Confluence

Let's talk about a problem worth solving.

Whether you're modernising a platform, building AI infrastructure, or just want to compare notes on product craft — I'm always up for a good conversation.

vasurajatkumar1993@gmail.com
Email
→
linkedin.com/in/vasurk
LinkedIn
→
+44 7586 596826
Phone
→