Tarushi Pathak Email and Phone Number

- Optimized and developed the capability to print Ads based on geographies in a service which prints 10 million shipping labels per day for Amazon deliveries across the world. The success of this pilot opened doors for Amazon to launch the project with a profit estimate of $8.9 million. - Developed a compression algorithm to reduce the payload size of a shipping label with Ads by ~40%, which helped adapt to low internet speed( around 15 kb/s ) at Amazon Fulfillment Centers, thereby… Show more - Optimized and developed the capability to print Ads based on geographies in a service which prints 10 million shipping labels per day for Amazon deliveries across the world. The success of this pilot opened doors for Amazon to launch the project with a profit estimate of $8.9 million. - Developed a compression algorithm to reduce the payload size of a shipping label with Ads by ~40%, which helped adapt to low internet speed( around 15 kb/s ) at Amazon Fulfillment Centers, thereby unblocking the pilot for the Ads on Labels project with a profit estimate of $8.9 million. - Developed a service responsible for pre-production testing for shipping labels use-case using AWS Lambda ( Java ) with its data-source as Amazon DocumentDB which streamlined testing, ensured robust label generation and thereby, minimizing deployment risks.As a part of this I also designed and developed a ranking algorithm which ranks the data received from the database for a given request.- Recognized the potential for efficiency gains in business intake request scoping by introducing a zero-cost and scalable solution, slashing scoping turnaround time by 95%. Implemented using React.js and AWS Lambda with Python, ensuring a seamless and responsive application.- Successfully mentored SDE Interns, guiding them to deliver impactful projects while fostering their professional growth and development.- Participated in on-call rotations and provided support for high severity issues for the service responsible for printing shipping labels which averages around ~10 million calls per day. Show less

- Leveraged Machine Learning to automatically identify font differences in generated regression reports, achieving a recall of 0.99. This significantly reduced the need for manual verification and accelerated the process by 80%.- Prepared a python script to identify characters which were leading to failure in label generation.- Worked and successfully launched multiple carriers into the Amazon transportation space as part of Carrier Integration team.

Completed the following research projects:1)Predicted UTS using Machine Learning2) Used DCGANs to generate microstructures 3) Did Grain Analysis of Microstructures through Image PreprocessingThe last two projects were published in journals Indian Journal of Data Mining and ADCAIJ respectively.

- Researched and developed 3 algorithms for autonomous navigation for an UAV along with my team.- Recruited and mentored new students in the field of IOT.

Researched and developed an algorithm for enabling navigation for a dual copter.