May 2022 – Jul 2024 · 2 yrs 2 mos

Sep 2020 – Apr 2022 · 1 yr 7 mos

• I have been in charge of a five-person team as an engineering manager. Primarily concentrating on product coordination, roadmap talks, cross-team project management, team building, and the management and personal development of team members.
• Have been actively involved in technical projects and have helped to develop new ones for system optimisation, infrastructure upgrades, and search relevance enhancements (Lyrics search, Phonetic handling, Named entity detection, etc.).
• DeepSearch -> created, tested, and released a deep learning neural network model for ranking personalised search results. Completed the work associated with
• 1) Defining the entire model architecture.
• 2) Selecting the pertinent metrics and loss.
• 3) Preparing and analysing training data.
• 4) Developing and perfecting several model instances.
• 5) Engaging the team members to productionize the prototype model
• 6) End to end automated training and deployment pipelines
• Migrating Search service from AWS to Azure -> Planning the migration strategy, supervising team members and addressing roadblocks, helping them finish tasks on time, and giving the team pertinent training to enhance their general understanding of the systems
• Infrastructure Optimization -> Developing a new architecture to optimise search relevance and latency for the SOLR cloud setup. primarily assisted the team with implementation details and worked on projects involving project design.
• Migrating Search service from Azure Singapore to Azure India -> primarily in charge of assisting team members with project planning, selecting smaller tasks to accomplish the main goal on schedule, assisting with cross-team coordination to resolve obstacles, ensuring zero downtime for users, and organising pertinent training.

Aug 2018 – Aug 2020 · 2 yrs

• Serving as the technical lead and acting engineering manager, I'm in charge of the whole search system and collaborate with the AI team to enhance recommendations.
• Topic Recommendation: By utilising Learning to Rank models, users may receive music recommendations based on a variety of user history-related characteristics. Accountabilities
• 1) Model choice and fine-tuning of parameters
• 2) Preparing data for training and testing
• 3) Front-end code modifications and modular deployment
• 4) Consistent updates and a daily forecast pipeline
• Search Optimisations: Determining and ranking infrastructure and search enhancements. Accountabilities
• 1) Enhancements to search features (such as entity identification, spell checking, and results ranking)
• 2) Infrastructure scalability and latency reductions
• 3) Infrastructure configuration, oversight, and upkeep
• 4) Managing a staff of six while serving as an interim manager
• 5) Assisting and directing the group on initiatives
• 6) Roadmap conversations

Jul 2015 – Jul 2018 · 3 yrs

• Search Relevance -> Enhancement and improvement of Search queries, improving cache settings, improving schema design and document ranking mechanism.
• Infrastructure -> Implementation and improvement of SolrCloud infrastructure,
• developing and improving data ingestion pipelines, Implementing, maintaining and improving micro service architecture using websockets, creating central search related metastore using dynamodb
• Monitoring -> Using apache spark to develop Search Click through rate (CTR) calculation algorithms,
• Developing search monitoring dashboard using cakephp,
• SolrCloud infrastructure monitoring, Websocket monitoring, System level optimizations.
• AI/Recommendation -> Currently working with Learning to rank models for user music topic personalization. Responsible for feature development, model design and data pipelines
• Technologies used -> qubole, spark, airflow, php ratchet, haproxy, sold, amazon ear, mysql, elk stack, mongoldb, zeromq, dynamo db, android studio (beginner)

Jul 2014 – Jun 2015 · 11 mos · Bangalore, India

• Development of Trace Replayer – Generate IO’s which will emulate an environment as if actual client is issuing the commands.
• •Worked on analyzing how different fields of a trace help in determining the behavior of the IOs.
• • Developed a model to segregate the IOs into different buckets based their properties.
• •Wrote scripts to feed the output of the model to SANBlaze which will finally generate the IOs.
• Technology used - SanBlaze, R language, filers
• Predicting the future value based on previous patterns using neural networks.
• Extensively experimented with the Xapian database, REST API's, and natural language processing. Determined the effect of description entered by the user in predicting the correct team the error is to be directed.

May 2012 – Jul 2012 · 2 mos · Mahan, Madhya Pradesh

• Development of Web Portal for Essar Township - Developed a portal to manage room allotment details.
• •Worked on the part related to accessing and editing the database.
• •Used normalization techniques to reduce the redundancy in database.