▪️ Working on a project on distributed machine learning with decentralized data in a mobile environment

▪️ Planned and led weekly sessions on C/C++ programming, GNU tools, data structures, and algorithms▪️ Wrote testing scripts for grading assignments with Google Test, Python, and Bash

▪️ Investigated federated learning approach to beam prediction problem in mmWave communication systems▪️ Proposed personalization and clustering techniques for handling data heterogeneity among base stations

▪️Worked on ML evaluation tools used for Ads personalization models in Facebook AI Infra team▪️Improved the evaluation time from days to hours using multithreading and PyTorch JIT scripting▪️Proposed and Implemented a tool that helps ML engineers to find an optimal quantization scheme

▪️ Designed and implemented an IoT device targeting system in Alexa Smart Home API▪️ Deployed the system to production using tools for continuous integration, unit testing, and A/B testing▪️ Participated in a hackathon to build Alexa custom skill connected with serverless web app using AWS services

▪️ Conducted individual research regarding parallel distributed machine learning (ML) in the edge network▪️ Developed a Java application that simulates computational cost and loss of distributed ML in edge nodes▪️ Proposed a scheme that optimizes ML in edge network by dynamically controlling the level of delay in terms of synchrony▪️ Accepted as a paper for the International Conference on Parallel and Distributed Processing Techniques and Applications

▪️Participated in a project for on-loading spatiotemporal context in mobile devices using differential privacy▪️Developed a front-end application for android implementation of contextual data on-loading architecture▪️Implemented SQLite database and Google Maps API for querying ‘Z context’ in contextual data