May 2021 – May 2024 · 3 yrs · Bangalore Urban, Karnataka, India

• ● Worked in the search platform team, handling a baseline load of 30K RPS and peak traffic of 100K RPS while managing large-scale distributed systems with Solr, RocksDB, HBase, Storm, Redis, and Couchbase. Worked with various microservices and designed new capabilities in the system while keeping scale in mind.
• ● Led multiple optimization efforts, reducing search system response time by 40% through rewriting sharding strategies, doing JFR analysis to detect low-level application bottlenecks, implementing automated cleanup of unused signals from the index – and several other performance enhancements.
• ● Optimized the core microservice bootstrap time by designing and implementing a RocksDB-based local caching system. Developed a custom RocksDB wrapper from scratch, reducing developer effort and enabling new use cases with features like a custom replication layer, schema validation, and partial update support.
• ● Built a high-performance feature store from scratch to improve search, enabling efficient storage and retrieval of real-time features at scale. Designed and optimized the system to handle 60K RPS on a single 16-core machine with P99 latency under 10ms.
• ● Led multiple business-critical projects to expand system capabilities, including the successful launch of Flipkart Minutes from the search infrastructure side, ensuring seamless launch.

May 2019 – May 2021 · 2 yrs

• ● Led the end-to-end development of various microservices along with a small team of developers.
• ● Took complete ownership of creating the low-level and high-level designs for various microservices like the User’s Journey, Gamification, and Market-intelligence system.
• ● Journey and Gamification systems were built with the scale in mind. A single consumer of the system ingested more than 400 events/sec at the peak time with an end-to-end latency of less than 1 sec. Overall, the system’s P99 was well below 20 ms at the throughput of 1K RPS at the peak time.
• ● Designed and developed a market intelligence tool to derive price intelligence daily. This was a Python-based project. The pipeline ingested Millions of catalogs every day from different pluggable sources and did various computations on them with a very minimum end-to-end lag.
• ● Took various tech initiatives in my pod-like setting up a complete monitoring system for the microservices using Prometheus and Grafana, doing perf at 10x of the current scale to eliminate any future bottleneck well before time, and documenting all the work to make the onboarding of the new teammates easier.

Nov 2018 – Apr 2019 · 5 mos

• ● Spearheaded the development of a cutting-edge price intelligence tool designed to extract product data from various e-commerce platforms. This tool empowered our team to generate comprehensive reports, highlighting product commonalities between Meesho and other platforms while analyzing price differentials across the market.
• ● Pioneered the creation of an in-house image matching system, revolutionizing the process of identifying comparable products within Meesho's inventory. This innovation streamlined the handling of out-of-stock items by seamlessly offering alternative options, a task previously carried out manually.

Nov 2017 – Nov 2018 · 1 yr · New Delhi Area, India

• ● Developed some core microservices enabling the push notification system.
• ● Developed Python libraries for the internal system.
• ● Worked on the scalability issues of the PHP monolith to bring the load time down by 30%
• ● Contributed to the in-house node framework used by all the microservices.

May 2017 – May 2017 · 0 mo · Remote

• Made enhancements to the website which is used to manage PyCon in India.
• Link: https://github.com/pythonindia/junction

Feb 2017 – Feb 2017 · 0 mo · Remote

• DuckDuckGo is currently focusing on becoming the best search engine for programmers by providing instant answers and solutions to their queries.
• Instant Answers appear on DuckDuckGo.com above ads and organic links, on millions of searches. Instant Answers are created by an open source community of developers around the world.
• An example query of my work is https://duckduckgo.com/?q=npm+install&atb=v47-2__&ia=web

Jun 2016 – Mar 2017 · 9 mos · Remote

• Links to some of the problems are:-
• Business work(https://doselect.com/playground/j6d5v/problem/l1q5a),
• 2012 Virus(https://doselect.com/playground/j6d5v/problem/84y8o),
• Basic networking(https://doselect.com/playground/k0nde/problem/yby58)

Jun 2016 – Jul 2016 · 1 mo · India