Oct 2022 – Present · 3 yrs 5 mos

• 1. Designed and implemented a scalable engineering infrastructure to deploy and run machine learning models for targeted ad placement in Prime Video. Optimized GPU compute choices, achieving a processing rate of 1,000 videos per day within 3 months. Collaborated closely with applied scientists to seamlessly integrate the models into the production pipeline, delivering a high-performance solution for US and EU markets.
• 2. Designed and implemented a compliance workflow for identifying Election misinformation and hateful content, this reduced average handle Time for Elections policy coverage by 50% from 90 min to 45 min
• 3. Extended Video Similarity Service to add newer algorithm of finding similar videos. Refractored existing service by adding improvements, creating common libraries, documenting onboarding steps which reduced time to onboard a new similarity algorithm more than 50%, reducing from 3 months to around a month.

Jul 2021 – Present · 4 yrs 8 mos

• 1. Worked closely with PMs and multiple teams, disambiguated the problem and came up with designed and implemented optimal assignment of tasks to DC associates.
• 2. Implemented end to end analytics feature for viewing near real time produce graded data in Distribution Centre in 2 months for a critical feature delivery.

Oct 2019 – Jun 2021 · 1 yr 8 mos

• 1. Created data Ingestion component of data lake. Made it configurable using CDK constructs and parameters making a new ingestion almost completely automatic, reducing dev time for onboarding from 2 weeks to a day.
• 2. Evaluated Project requirements and on-boarded multiple teams into data lake and helped them structure their schema.
• 3. One of our systems listening to event notification, during backfill due to storm of messages, took days to process messages. Introduced Deduping solution which improved latency by reducing multiple processing of duplicate messages, around 39.39 % notifications (12,82,966) notifications were removed
• 4. Performed benchmarking and load testing of Reporting services to 2x times the expected trafiic for Peak times. Identified and proposed solutions to improve performance, which reduced lambda processing time by 1/5th.
• 5. On-boarded multiple SDEs into team and mentored an intern to deliver her Intern project.