Oct 2021 – Dec 2023 · 2 yrs 2 mos

• Led development for building an internal vector search database named Hamsa.
• Eventually consistent, fault-tolerant distributed vector search database. A cloud-native database built on top of Kubernetes, FileStore, GCS, and Tensorflow Serving. Supports various index-building algorithms. Configurable index-building frequencies.
• Used for diverse use cases: candidate generation, duplication detection, cold start in feed, etc.

Apr 2020 – Oct 2021 · 1 yr 6 mos

• Led implementation of multiple components for the launch of our short video app Moj (a competitor of TikTok).
• Led implementations of Feed Delivery and Ranking systems, scaling the video feed from 20 million to 180 million monthly active users. These components are related to feed delivery, e.g. feed-relevance-services that can fetch various candidates, rank them, and apply post-processing layers like diversity, filtering, etc.
• Worked in building and scaling real-time FFM training system.

Mar 2019 – Mar 2020 · 1 yr

Oct 2018 – Mar 2019 · 5 mos

• Led development of events platform processing hundreds of Billion events daily. This included setting up of end to end event sourcing, dynamic schema updates, and various ETL jobs written in Google Cloud Dataflow to consume from various message queue systems like Google PubSub, Amazon Kinesis, Apache Kafka to data lake Google BigQuery. Database trigger-based updates to maintain fact tables.
• Worked on the efforts of splitting the common backend monolith into splitting various microservices.
• Implemented ML models for retrieval: Based on Popularity and also Content-Based and Collaborative Filtering candidate generators.

Jan 2017 – Jan 2018 · 1 yr · Mumbai Area, India · On-site

• Worked with [CodeChef](www.codechef.com) as a Software Engineer.
• Developed various backend features in PHP.
• Developed a high-scale ranking and rating system for programming contests. The rating system was based on ELO based mechanism.
• Worked on optimizing performance bottlenecks in the MySQL database for the discussion platform.

Aug 2015 – Jun 2017 · 1 yr 10 mos · Bengaluru Area, India · On-site

• During my journey of approximately 2 years on LinkedIn, I worked as a backend software developer majorly on the payments team and the mobile team. The work involved technologies related to Java/Scala backend systems, Android Systems, and Mobile technologies. Also worked with big data platform systems on top of the Hadoop file system, Spark, Kafka, and Samza.
• Created frameworks for handling recurring chargebacks for various payment gateways as a part of the Payments team.
• Implemented various web dev-related features. Built chrome notifications on the mobile website (LinkedIn Lite/Bolt project).
• Handled registrations/logins for all the phone-only users. Used Hadoop workflows for analyzing the registration and login funnel of the users. Created pipelines for
• sending SMSes to users at various stages of the user life cycle.
• Developed tools for the user of Customer Service Operations to help resolve the payment-related issues related to users.
• Built multiple notification and SMS campaigns to increase user engagement and retention on the mobile website. User targeting was done using churn prediction models.

May 2013 – Jul 2013 · 2 mos

• Implemented MS-SSTP VPN protocol implementation in the Android platform. The protocol is primarily used to establish a secure VPN connection from a mobile (Android) device for accessing a remote private network over a public network. Used Native JDK(Java Development Toolkit) for integrating native C++ codebase with Java in Android Systems. Was offered a Pre Placement Offer for the work.