Jun 2022 – Nov 2023 · 1 yr 5 mos · Bengaluru, Karnataka, India

• Designed & led the development of content moderation systems to identify and act upon misinformation, impersonation and pornography.
• Link: https://www.livemint.com/technology/apps/koo-introduces-new-safety-features-for-content-moderation-11679567037371.html
• Benchmarked various LLMs and used it to be a single solution for all Koo Enrichment processes.
• Led the development of creator tools using GenAI Tools.
• Link: https://www.businesstoday.in/tech-today/trending/story/as-twitter-struggles-under-musk-its-indian-rival-koo-sees-an-opportunity-introduces-new-features-352611-2022-11-11
• Architected and led the efforts to develop a system to identify trending news on the platform.
• Led the efforts to make 'Topics' live in 10+ languages. Link: https://www.businessworld.in/article/Koo-Launches-Topics-In-10-Languages/24-08-2022-443475/
• Extended topics to multimedia content (images/video) as well.
• Ideated, architected and led the implementation of ‘For you tab’ a hub for personalised recommendations based on user activity patterns.
• Architected a central ML pipeline to run all ML-based enrichment modules in a scalable fashion.
• Architected a search indexing pipeline to aid in better search experience and relevancy of results.
• Architected an adaptive tagging pipeline.

Apr 2021 – Jun 2022 · 1 yr 2 mos

• Ideated, architected and led the launch of a real-time personalization system, which makes the users' feed more reactive to interaction leading to a 93% increase in engagements.

Sep 2020 – Mar 2021 · 6 mos

• Built a recommendation system that performed ~30% over the existing content delivery system
• Led the creation of a profanity detector to detect obscenities across video, text and voice.
• Spearheaded a solution to the cold-start problem of recommendations which led to eliminating 16 man-hours a day.
• Led the building of multiple attribute predictors to aid in data availability to various other systems.
• Created and set up systems to aid in automation, monitoring and optimisations across data analytics, data
• science and data engineering use cases.

Apr 2019 – Sep 2020 · 1 yr 5 mos

• Responsible for various exciting projects pertaining to data & distributed computing.
• Notable Projects: RentoBot, RentoOCR, AutoKYC, Invoice Generation System, Recommender System, RentoSizing, Data Collection & Scoring.
• Built a face detector to verify the genuineness of customers.

Jan 2018 – Apr 2019 · 1 yr 3 mos

• Prototyped a conversational bot to create newer experiences for users as well as taking load off the IVRS system.
• Built a 2 stage recommender system to aid in better product discoverability.
• Built a system to extract key information from govt. documents submitted by the user to aid in faster KYC.
• Created credit profiles of users to estimate risk. 
Link: https://techcircle.vccircle.com/2018/04/25/rentomojo-is-using-machine-learning-to-create-credit-profiles-of-users
• Rebuilt the existing invoice generation system to reduce the time taken from ~ 5 – 10 hours to ~10 minutes.
• Built a brand-agnostic sizing system for apparels by making use of data clustering and labeling the clusters.
• Created pipelines for collecting data from various sources to aid in the smooth functioning of several teams & speed up business processes. e.g. one of the pipelines saves ~750 minutes per day for an entire team. These qualitative data was converted to quantitative ones so as to automate decision making.

Apr 2017 – Apr 2017 · 0 mo · Ministry of Civil Aviation

• Was responsible for mentoring the team, ideating, building and presenting features.

Nov 2015 – Dec 2017 · 2 yrs 1 mo · Palo Alto

• Was responsible for building the entire data pipeline of the company.
• Notable project : Data Parser & Analysis.
• Featured as a maker in ProductHunt twice and trended in three categories (AI, bots, popular) at ProductHunt.
• Built a parser which would analyze a summary of text and categorize it to the best fit domain using word2vec.
• Gathered, restructured and stored historical research grants and patent data for fast retrieval from ElasticSearch.
• Analyzed $300bn+ worth of US Govt. data to provide insights about schools & research fields.
Link: https://www.producthunt.com/posts/senseihub
• Developed an algorithm to extract key concepts and built searchable topic maps using it, by analyzing news articles, to provide valuable insights. Link: https://www.producthunt.com/posts/topic-graphs-by-senseihub

Jun 2015 – Jul 2015 · 1 mo · Kolkata

• I was mainly responsible for the development of the Bidder Module and validations for an e-Auctioning Portal, NERAMAC1, commissioned by the Govt. of India.

Jun 2014 – May 2016 · 1 yr 11 mos

• Got selected for Mozilla’s “Summer of Contributions” Program and was the only one to successfully complete it.
• Enlisted in the official team roster of ‘The A-Team’ at Mozilla and added to the official credit list of Mozilla. Link: https://wiki.mozilla.org/Auto-tools/The_Ateam
• Listed as a contact for developers contributing to AlertManager. Link: https://wiki.mozilla.org/Auto-tools/Projects/AlertManager#Getting_Started
• Calculated & visualized metrics which would be used to track the progress of development.
• Reduced loading time from 5 seconds to 50 milliseconds.
• Automated the process of calculating improvements and regressions.
• Handled changed preferences and introduced finer control over the behavior of tests.
• Enabled test isolation in Mozilla's main test harness by running each set of tests in their respective directories in a single instance of the browser.