Working with Ads bidding Infra team to build reliable and scalable infra for Ads pacing products

• Led the design and launch of Bidding Oracle- a scalable microservice offloading ML computation from Ads pacing service, mitigating 5× YoY memory growth, unblocking 1%+ long-term iRev and improving reliability of the Ads pacing service by 27%

• Infra based improvements in Ads Retrieval -Owned and delivered a multi-year infra strategy to unify and improve the quality of pacing data across Ads indexing and retrieval, enabling 0.11% short-term iRev

• Ads Pacing reliability Led the design and implementation of a low-latency ads pacing reliability and backlog detection system, detecting delivery failures, retrieval/ranking blockages, and spend anomalies to prevent $70M/year in revenue fluctuations

• Designed and developed an advisor framework on the NOW platform, which enables content creators to develop targeted workflows specific to a user/organization

• Created a Skills prediction plugin in workforce optimisation suite for ITSM, that recommends skills needed to resolve open tasks. Trained a multi-label classifier for this purpose with an avg. accuracy of 76%.

• Worked with Oracle’s Commerce Cloud teams to develop features for the agent application of the Commerce platform.

• Developed functionality for placing return/exchange requests, integrating tax processing engine with the shopping cart and configuring wish lists by B2B enterprise’s customers. Created RESTful APIs using JAVA for the backend.