🧠 Defined and led product strategy for Snap’s AI/ML Platform, Bento, a central platform enabling ML Engineers to power Ads, Content and Growth Recommendation Systems (RecSys) including Ad Ranking, Stories, Spotlight, and Friending. Built roadmap for 100+ engineers to deliver Bento's Feature Engineering, Training Data, Model Training, Inference, and ML Observability sub-platforms, leveraging open-source AI Frameworks like PyTorch, TensorFlow, Flink. Expanded the platform to support GenAI use cases including hosting open source LLMs like DeepSeek for GenAI-powered content understanding use cases. Defined the next evolution of the platform around Agentic AI, enabling autonomous end-to-end ML development from exploring hypotheses to training, evaluating, and proposing models for production.

🎯 User Age Prediction Product: Envisioned Snap’s ML-based User Age Prediction product, bootstrapped a cross-functional team spanning data science, engineering, legal, program and product, and shipped a solution predicting age for 450M+ daily active users.

⚡ Snapchat Lite: Improved Snapchat’s unit economics by taking an ambiguous cost-reduction mandate and assembling a cross-functional team across different product groups (Ads, Content) and disciplines ( data science, MLEs, engineers, UI, Designers). Identified high-impact projects focused on low-monetized/low-engaged users, and delivered $xxM in savings, by taking deliberate tradeoff across cost, engagement and revenue.

PM in Meta's AI Platform team, with special focus on following 3 products:

1- Facebook Feature Framework (F3) and Feature Store: Led Meta’s Feature Engineering platform from ground up for all AI-heavy teams (Ads, Reels, Feeds, Marketplace etc.). This efficient and developer friendly system helped Meta deliver major business wins (e.g. Ad Score gains, Reels Watch-Time improvement) by supporting development of high quality realtime and batch features. It also saved thousands of engineering hours, by consolidating highly fragmented ML stack of Meta on a single efficient solution, by encouraging feature sharing, providing resource optimization and friendly DevX.

2- AI Metadata Platform: Led AI platform that builds AI Knowledge Graph for Meta. The platform builds an inventory of all AI artifacts (Models, Features, Datasets, Labels, Predictions etc.), builds lineage across them (e.g. does this Ads model, use that dataset), and collects additional metadata (owners, retraining schedule etc.). This platform supports AI use cases around Governance, Efficiency, and Compliance.

3- AI Privacy: Worked on the AI platform that makes Meta's ML stack privacy compliant. It ensures adherence to current and future regulations. This included US Compliance: FTC, CCPA, CPRA, USO

EU Compliance: GDPR, ePD, IDPC, FCO. Building this platform helped build user trust, mitigate potential lawsuits resulting in savings of billions of dollars, and saving thousands of developer hours by building privacy solutions centrally.

Product Manager for Bing's syndication business, a product that handles 13%+ of search engine market share (on top of Bing's own 25%) in US desktop market, per Comscore 2019 report. This is one of the biggest search marketplace on the planet with publishers from the likes of Bing, Verizon Media, Ecosia, DuckDuckGo. I am the lead PM for this product, and I envision, ideate and handle execution of different Bing API offerings to provide state of the art search capability to these publishers. I focus on increasing revenue, customer acquisition and customer retention and work towards finding the right product market fit for Bing APIs.

This is a position in Bing & Cortana Ecosystem team at Microsoft. In this role, I manage Bing APIs and work towards increasing Bing market share and revenue, by integrating Bing with various products within and outside Microsoft. An example of Bing integration is, Bing powering Yahoo and AOL search results. Bing is also bringing the knowledge of web to various products like Sticky Notes, Outlook etc.

This role is unique as it is a good mix of TPM and Product Management with a flavor of Business development.

In this role, I owned the complete Program Management of the projects mentioned below. PM role expects me to envision the product/features, do analysis and do competitive study to support the product/features that I am recommending. I also design the functional specification of the product and work on its execution along with my developer and test counter parts. I am also responsible for making sure the product ships smoothly and is maintained once it is shipped.

1. BING LOCAL: Defined product strategy and managed Bing Local Data product team that powers the search for local businesses for 100s of millions of monthly queries across Cortana, Maps, Windows, Office and many other Microsoft canvas. Managed Bing Local Data Platform, a machine learning based engine responsible for ingesting, web mining, crawling, matching, merging and publishing millions of places and local business listings in the US market and global markets

2. RANKER SHIPPING PLATFORM: Worked as the Technical PM for the platform that programmatically tests all Bing Rankers for highest quality and reliability and against various metrics. Platform also makes sure ranker has all required meta streams to run smoothly and handles the deployment of ranker in all production environments.

3. WEB RANKING: Improving the relevance and ranking of web results for “health” segment. This project expected me to have good machine learning understanding.

4. STRUCTURED DATA EXTRACTION: Technical PM for the product that extracts useful information from the websites. I also managed a team of 9 people, who wrote models for this data extraction.