May 2022 – Aug 2022 · 3 mos · San Francisco, California, United States

• Engineered a Python service running as Kubernetes CronJob to expose continuous delivery pipeline usage metrics via dashboard and REST API to detect deviations from supported patterns, track growth and plan roadmap projects.

Jan 2022 – Dec 2022 · 11 mos · College Station, Texas, United States

• The research work aims to find solutions to reduce data quality issues for surveys hosted on crowdsourcing platforms like Amazon Mturk. The data quality can be significantly improved by identifying random responses, intentional trolling, satisficing, and bot responses in online surveys using Machine Learning. Detection of bad responses with high confidence means better data quality leading to accurate research results.

Feb 2019 – Jul 2021 · 2 yrs 5 mos

• Worked on multiple projects in GoFood, one of the largest food delivery businesses in the world.
• 1. Search and Recommendation Features
• End-to-end ownership of features to improve and personalize customer's search and recommendation experience.
• Generated cart recommendations on merchant profile page based on frequently ordered together items using the industry leveraged Apriori algorithm resulted in a drop in time to order and increase in order value.
• Engineered the first version of query understanding to classify the search query's intent and suggest relevant searches or results based on each identified intent.
• Constructed restaurant search history using Kafka and Redis helped users get to their recently and frequently viewed restaurants faster and easily.
• Upgraded the static top searches so that they are contextual, more granular by area, dynamic, and intent embedded, powered by Elasticsearch and Kafka.
• Build data pipelines for customer and restaurant features for search personalization using PySpark, Airflow, Kafka, BigQuery and GCS.
• 2. Reliable and scalable search systems
• Increased system uptime, reduced latency, and decrease the infrastructure costs by multiple initiatives:
• Upgraded Elasticsearch cluster from v6.3 to v7.5 with zero downtime. Tuned garbage collector and heap size.
• Location-based sharding of the restaurant and menu item indexes.
• Profiled and optimized slow queries.
• Batch updates to elasticsearch.
• Created an elasticsearch load testing framework for benchmarking.

Jul 2017 – Jan 2019 · 1 yr 6 mos

• 1. Monolith to Microservices
• Helped decouple systems and team to ship features faster and reliably.
• Split Ruby on Rails monolith service into content and search microservices written in Golang to scale teams to make features get shipped faster, faster development feedback and less error-prone.
• Lead the effort to upgrade the PostgreSQL database to the latest version with zero downtime.
• Decoupled the services by event-driven architecture using Kafka.
• 2. Refactored search service to a single backend cluster to serve traffic from all the countries decreased infra cost by ~30% and maintenance overhead.

May 2016 – Jun 2016 · 1 mo · Bengaluru, Karnataka, India

• Rewrote a critical component in C++. The component was part of a legacy platform to enter, manage and execute the firm’s forex transactions. The code was successfully serving the production traffic at the end of the internship.

May 2014 – May 2014 · 0 mo · Greater Indore Area