Designed, coded, and managed a platform to schedule notifications for over 100 million customers.Supported multiple A/B tests targeting different customer segments simultaneously, using configurable AI models.Implemented the scheduling system using PySpark.Developed and integrated in-house algorithms for enhanced functionality.

SWE on Gotham Search

CA - advisor for COMS 4113 Distributed Systems

Head Teaching assistant for EECS 4750 Heterogeneous Computing:* Created assignments and exam for the course for programming in PyOpenCL and PyCuda covering concepts like, tiling, shared memory, constant memory optimizations. * Held office hours and evaluated assignments, projects and exam.

Heterogeneous Computing Team.Worked on optimising Lidar and Perception algorithms for GPU. Learnt Software development in C++ using CMake, Jenkins, GTest and other required frameworks. Also learnt about CUDA and optimising algorithms using techniques like Tiling, Shared Memory and Constant Memory. Work was majorly in parallelising bunch of for loops and Matrix Multiplication and Transpose based algorithms. Also worked on optimising some OpenCV functions.

Working as a Teaching Assistant for Programming Language Technologies.

Worked as an SDE for E-HVA Team, Jan 2020 - Dec 2020: Worked in PySpark, to refactor and redesign 10000+ lines of code written by ML Scientists on Hidden Markov Model. Improved the worstcase runtime of the feature-generation by ~80% also saving on memory requirements by reducing shuffle writes by 75%, using a custom bitset based count distinct algorithm that I designed, this helped save $10000+ for the team per year. Improved runtime of training, scoring, pre-processing and postprocessing, by ~30% by optimizing on caching, partitions, and joins Currently working on automating on-boarding process for our customers. Worked as an SDE for Data Ingestion, Jan 2019 - December 2019: Designed, created and managed a java based service for scheduling jobs according to a cron schedule for big data processing.I created several APIs for CRUD and operational readiness. Used AWS DynamoDB as a database. Created a UI over Kibana using metadata from dynamoDB streams and indexed the metadata to AWS Elastic Search using AWS Lambda. This scheduler used to schedule 5000+ schedules daily with a availability rate of over 99.99%.Worked as an SDE for Asin Data Service Team July 2018 - January 2019: Asin Data Service is a service on thousands of discrete applications running in a distributed environment that ensures that data from more than 40 services is aggregated for display on Amazon’s websites – 10s of thousands of times every second with minimal latency. A bad deployment to AsinDataService used to result in bad customer experience for atleast 4 hours in North Amerca region and an average of 2 hours in other region, in addition to this the team was also experiencing several issues during deployment(~High severity issue, sev-2 at Amazon) for which there was no root cause, it turned out that both were inter-related. I root caused the issue to JVM warmup. After the fix I put in place the deployment time reduced by 35+% and Number of issues dropped down to 0, from 22 per week.

Teaching assistant for:• Introduction to programing in C• Advanced Data Structures and AlgorithmsResponsibilities: Grade exams and assignments, conduct quizzes, conduct office hours to clear doubts.

Worked for Giveaway team: I was tasked with adding a new requirement for eligibility to Amazon Giveaway. I created seller side webpage "enter to giveaway with requirement as subscribe to newsletter". I created another customer side page where customer had to fulfill the requirement of subscribing to newsletter to be eligible for giveaway. For this project I worked on JAVA, HTML, JAVASCRIPT, XML, SPRING. Timely delivery and deployment to production by the end of my internship helped me get pre placement offer at Amazon.

Key highlights:• A member of organizing committee for Megathon (Largest student lead hackathon in South India) 2017.Other Responsibilities: • Conducting sessions, inviting speakers to promote startup culture within Campus.