PS: We are hiring.

Working as a full stack production engineer in the DCOI (Data Center Orchestration and Intelligence) team to help manage and scale the AI infrastructure at Meta.

Sensor simulation and synthetic data at large scale. worked with research and engineering teams on production of high-fidelity datasets for developing, training, and validating state of the art computer vision and machine learning systems.* Performance optimizations for Unreal Engine rendering pipeline* Customization of Ground truth modalities* Creation of internal self-serve application for generating synthetic datasets (full-stack)* Architecture and implementation of a distributed production system* Design, setup, and maintenance for a custom internal render farm to support production needs

Worked on various features for launching next generation of Microsoft Store.Build various Platform Clients in M365 Core-Substrate for ESP services which are at scale of 100’s of billion requests per day.

Ant is a chat based AI Model, Acquired by Snap Inc.(Known as Snapchat) in Oct. 2021

Worked in Content Ranking/Personalisation Team.Used Apache Beam/Dataflow to create pipeline that output Discover Feed training data to BigQuery (processed ~ 40M examples/day, wrote filters/joins to store 800-900K inputs w/ representative features and labels).Built suite of data analysis functionality for EDA, production model evaluation, etc. w/ Pandas, Jupyter, IPython Widgets, BigQuery client, Seaborn (wrapper functions and UI to slice, add ratio fields, visualize data, ect.)

Worked on backend and infrastructure projects for the Low Voltage Software team.Received strong hire full-time return offer.

Worked on backend projects on the IVS (Interactive Video Service) Core team, which bridges the gap between the Twitch's Video Platform and AWS.

Worked on Apple Podcasts team.

I worked for the Infradev team, Production and Infrastructure Engineering.- During my internship, I developed a RealTime UI to demonstrate the real-time traffic between sites/services, etc.- The traffic data was stored in Kafka and it was fetched using consumer application coded in Flask.- Redis was used to share the message to the WebSocket connection.- UI was implemented using Cytoscape.js.- Tech Stack Used: Kafka, Flask, Redis, Socket Io, and React.Received strong hire full-time return offer.

Part of Azure Compute United States Data Science Team.Created a tool for automation of reports which was integrated as a part of an internal project at microsoft.Received strong hire full-time return offer.