Oct 2016 – Mar 2022 · 5 yrs 5 mos · Pune, Maharashtra, India

• I have identified opportunities, built and scaled several teams and have also introduced several new technologies and engineering practices at MindTickle.
• Joined MindTickle to develop product features in machine learning as a Product Researcher. Conceived, built and scaled analytics platform and team ground up. Developed few other features like natural language interface, transcription, automatic evaluation of practice sales pitches, auto-generated insights from analytics in natural language. Built and scaled Call.AI conversation intelligence product and team based on completely new architecture and engineering practices. Helped to reorganise the engineering into domain driven teams and transitioning teams to new leadership. Took over Devops functions and introduced infrastructure as code across all infrastructure and telemetry. Took over storage infrastructure team and brought down RTO / RPO to meet compliances, point in time disaster recovery, access management, storage access layer, etc. among the key few projects. Scaled both the teams and established the Site Reliability Engineering team. Let frontend infrastructure team to bring in Relay, GraphQL, micro frontend, architecture. Brought in technologies like Spark, Snowplow, React JS, Helm, Kotlin MP, Rabbit MQ, etc. Oversaw implementation of Datadog, Sumologic, Gitlab, Snyk, etc. Took over a major infrastructure upgrade project (for our user management) midway which had gone south and course corrected it, oversaw its completion while maintaining and optimising the older system in parallel and ensuring a smooth transition and site wide migration. Also had led mobile team briefly in crisis and ensured the delivery of the new mobile architecture while restaffing the team from scratch. Led program management function by transitioning to agile and JIRA advanced across all teams incrementally and set up the program management office by defining and hiring across various levels.

Dec 2015 – Jul 2016 · 7 mos · Bangalore

• Pin Click is a highly successful real estate brokerage firm based in Bangalore and Pune. We started up to bring managed rental accommodations and end to end property management as a service based on the deep understanding of the real estate market combined with AI and engineering skills which the founders collectively bring on table.

Mar 2015 – Dec 2015 · 9 mos · Bangalore

• SearchX was started up with the thesis that keyword search is not good enough in exploratory search, especially when documents are large. Such scenarios are common in finding legal precedences, prior art patents, academic and medical research and searching news, blogs, literature, etc. Based on my research in nearest neighbour search in high dimensions, SearchX aimed to bring a different method of searching and exploring information in such scenarios. You are more than welcome to reach out to me for more details.

Mar 2014 – Mar 2015 · 1 yr · Bangalore

• Worked on aspect extraction from product and travel reviews.

May 2010 – Mar 2013 · 2 yrs 10 mos

• Worked on spelling correction, auto complete and search suggestions in email search using learning to rank. Worked on application of Canonical Correlation Analysis in text search and mining comparable corpora and also co-developed a new technique for learning multilingual projections using dictionaries.

Mar 2013 – Feb 2014 · 11 mos · New Delhi

• Recruited on basis of 424th Rank secured in All India Civil Services Examinition conducted by Union Public Service Commission. It is an administrative service for Union Territories under the Ministry of Home Affairs.
• Worked as a probationary officer in training in various departments of Delhi Government and was trained in law, public administration and government office procedures as a part of the probation

May 2009 – Oct 2009 · 5 mos

• Worked on application of Spectral Partitioning techniques in Pseudo Relevance Feedback in Information Retrieval. Also co-developed a method for extracting ad keywords from URLs and web pages

May 2008 – Jul 2008 · 2 mos

• Worked under Raghavendra Udupa and A. Kumaran to develop and implement a language independent statistical model for mining named entity equivalents from an article in resource poor language using a similar article in resource rich language for which a named entity tagger is available
• Implemented several Hidden Markov Model based alignment models, increasing in complex-
• ity, for transliteration
• Implemented bigram language models and studied the eect of dierent smoothing tech-
• niques on them
• Implemented Support Vector Machine based discriminative transliteration models
• Proposed and implemented a mixture model comprising of transliteration, translation and
• acronyms to identify equivalent multi-word named entities in two sentences of dierent
• language
• Proposed and implemented a variant of viterbi algorithm to do the matching
• Did a thorough literature survey in transliteration, language models and smoothing tech-
• niques, generative models, existing techniques for single word named entity mining and
• other mining and boot strapping approaches used for CLIR

May 2007 – Jul 2007 · 2 mos

• Worked under Dunja Mladenic and Marko Grobelnik on automatically mining of named entities and their relationships from Wikipedia articles using machine learning techniques.
• Designed regular expression based parsers for extracting information from Wikipedia articles
• written in native markup language
• Wrote heuristic functions for labeling Wikipedia articles as person, place or organization
• based on extracted information with very high accuracy
• Used this labeled data to train Support Vector Machine based classier and hence classied
• rest unlabelled articles
• Mined dates and sentences containing them from these articles
• Used cross references between articles to identify inter-relationships between persons