Vamsi Krishna Varanasi
Summary
Principal Product Manager with 10+ years building platform, fraud, and marketplace products at Delivery Hero, Careem, and Ola. Hands-on AI builder shipping RAG, MCP, and agent workflows on GitHub alongside leading maps and fraud platforms at global scale.
Experience
-
Career break — AI exploration & side projects
Aug 2025 – Present
AI Product Builder / Career Transition
- Repositioning for senior product leadership in the AI era: refreshed identity across LinkedIn, resume, GitHub, and personal website.
- Running a listening tour with senior PMs and hiring leaders to understand what "good" looks like for AI-era product roles.
- Shipping hands-on AI projects (Vector Explore, Track Slack, Just Talk) to demonstrate RAG, MCP, and agent workflows end-to-end.
- Operating Sorted Signals, an n8n-powered AI newsletter that grew to ~200 subscribers and ~5,000 article impressions.
-
Delivery Hero SE
Nov 2022 – Aug 2025
Principal Product Manager — Incentive Fraud & Fraud Monitoring
- Owned the company-wide incentive fraud capability end-to-end (ideation, research, PRD, technical design, stakeholder alignment, development, QA) protecting an €890M/year voucher and discount spend.
- Shipped €260k/year in fraud savings across a limited rollout (3 of 65 countries), with the methodology and instrumentation ready for broader rollout.
- 1.95B fraud decisions served in 2024 across the global marketplace.
- Built automated feature cleanups, versioned features, and a live dashboard for monitoring feature health; surfaced real user needs by interviewing data scientists, ML engineers, and operations.
- Saved ~€456k/year in cloud and infrastructure costs by systematically reviewing RTM pipelines and eliminating inefficiencies.
- Mentored PMs and improved the product function by templatizing PRDs and A/B tests and upholding standards.
-
Delivery Hero SE
Mar 2021 – Oct 2022
Senior Product Manager — Maps & Location Platform
- Saved €232k/year on maps APIs by bringing in lower-cost providers and a zero-cost in-house solution (12% of maps cost).
- Reduced empty-address errors by 85%, improving conversion in address entry and ordering flows.
- Delivered a +25pp increase in address completeness via map-provider blending and data enrichment.
- Increased address search quality by 15% in Berlin with an experimental search-ranking algorithm.
- Curated geo-knowledge by building a database of 35,000+ tagged polygons crowdsourced from local teams.
- Built self-service tools that empowered local teams to manage their own geodata, reducing organizational friction.
- Evangelized the platform internally; onboarded 6 new internal customers.
- Hired and mentored 2 direct reports; interviewed 50+ candidates across functions and seniority levels.
-
Careem Networks GmbH
Apr 2020 – Feb 2021
Product Manager II — Maps & ETA
- Saved $200k/year by building a global trip reporting system and switching to transaction-based billing for Google Maps.
- Contributed to $3.4M/year in company-level savings by switching ETA monitoring from New Relic to an in-house APM product.
-
ANI Technologies Pvt. Ltd
Apr 2017 – Dec 2019
Senior Product Manager — Ola Cabs (Oct 2018 – Dec 2019)
- Reduced contact rate by 14%, achieving ~120k AUD/year in cost savings for Ola Cabs Australia.
- Built a support request history feature that increased customer satisfaction by 5pp and rating response rate by 11pp.
- Reduced peak-time in-trip ride cancellations by 9% via ride-allocation algorithm improvements.
- Grew Ola Postpaid to 300,000+ users at 0 customer acquisition cost.
Product Manager — Ola Financial Services (Apr 2017 – Oct 2018)
- Reduced 3rd-party fraud by 90% and prevented ~15k support tickets per month by building a fraud control system for Ola Money.
- Eliminated losses of ₹1M/day for Ola Cabs with a balance hold feature on the Ola Money fraud control system.
-
Securifi Embedded Systems Pvt. Ltd
May 2016 – Apr 2017
Product Manager, Usability
- Improved app usability of the Almond home IoT system by building coach marks, an in-app help center, and natural-language I/O for settings (NLG/NLP).
- Reduced getting-started support tickets by 70%.
- Created a new usage mode by integrating with IFTTT; grew the service to 150,000+ applet runs and 1,200+ users within 3–4 months.
-
Infosys Ltd
Jul 2010 – Mar 2015
Technology Analyst, CommerceEdge
- Built 20+ customer sales demos for a storefront-to-warehouse B2B SaaS platform with $10M sales potential.
- Cut demo turnaround from 14 days to 5 days (-65%) by standardizing the demo pipeline.
- Compiled customer feedback from demos into a prioritized feature backlog for the development team.
Selected achievements & impact
- Owned the €890M/year incentive fraud capability at Delivery Hero; shipped €260k/year in savings in a limited rollout (3 of 65 countries); 1.95B fraud decisions in 2024.
- ~€456k/year in cloud and infrastructure savings from systematic RTM pipeline review.
- €232k/year in maps API savings and 12% maps cost cut; 85% fewer empty-address errors; +25pp address completeness; 35,000+ geo polygons crowdsourced.
- Saved $200k/year (Careem) and contributed to $3.4M/year (Careem) in cost optimizations.
- 90% reduction in 3rd-party fraud; prevented ~15k support tickets/month; eliminated ₹1M/day in wallet losses (Ola Money).
- 6 internal customers onboarded to the maps and location platform at Delivery Hero.
Beyond the work
- Builds and shares reusable AI workflows for product managers — Vector Explore (RAG), Track Slack (MCP), Just Talk (agent skills) — on GitHub.
- Active listener: runs a continuing listening tour with senior PMs and hiring leaders on what good looks like for AI-era product roles.
Education
- Apr 2015 – Apr 2016
Post Graduate Program in Management (equivalent to MBA) · Marketing & IT
Indian School of Business, Hyderabad
- Sep 2006 – May 2010
Bachelor of Technology, Computer Science
Jawaharlal Nehru Technological University, Hyderabad