Mar 2025 – Jul 2025 · 4 mos

May 2022 – Apr 2025 · 2 yrs 11 mos · London Area, United Kingdom

• # Integrity
• Led and executed a collaborative strategic initiative to identify high-severity, low-prevalence harms in trust-and-safety context, implementing robust enforcement mechanisms through advanced ML techniques.
• Engineered prevalence measurement system for accurate evaluation of harm landscapes, ensuring compliance with legal standards.
• Pioneered deep learning methods to develop embeddings for text, image, and audio-visual modalities, enhancing classifier capabilities and driving innovation in the field.
• # Ads Ranking
• Led and executed strategic workstreams to enhance ads ranking algorithms, optimizing for long-term value and revenue by leveraging user and ad attributes, resulting in ads revenue lift.
• Applied ML techniques to isolate signal impacts and maximize long-term value using short-term information. Conducted extensive A/B testing to inform data-driven product decisions and enhance overall performance and user engagement.

Jun 2021 – Jan 2022 · 7 mos

• Led development of predictive system for health-risks for elective surgeries, improving financial readiness of customers.
• Developed Conversational AI System to augment support agents, reducing handling time by 25%.
• Created Long Format Document Search Utility using Elasticsearch, streamlining information search process.

Feb 2018 – May 2021 · 3 yrs 3 mos

• Data-driven healthcare ecosystem to make participants financially ready and health-risk aware of their elective surgeries & preventive medical conditions using medical claims.
• Best in academia and industry joint replacement risk prediction engine; ​Detailed work in the publication.
• Conversational AI System for Health & Wellness participants

May 2017 – Jan 2018 · 8 mos

• ● Health claim adjudication system using claims and user information for a leading insurance company.
• ● Conversational AI (chatbot) system for a travel aggregator and ticketing platform. Actions: Discovery, booking, cancellation, FAQ, user info, user booking info, alerts etc.
• ● Aspect-based sentiment analysis system for a banking client.
• ● Information extraction from KYC documents captured from handheld devices. Includes affine transformation, text segmentation and recognition.
• ● Fuel and time prediction simulator using 100's of flight sensor temporal information for a leading aircraft manufacturer

Jun 2014 – Apr 2017 · 2 yrs 10 mos

• Semantic search platform for scientific document discovery.
• Developed word hierarchy using word embedding and chronological information.
• For contextual search developed word embedding system; used LSH for nearest vector search.
• Developed conversational system to execute client specific tasks using user information and other details.
• Named Entity Recognition
• Intent recognition
• Word embedding for classification

Feb 2014 – May 2014 · 3 mos · Chennai Area, India

Oct 2013 – Jan 2014 · 3 mos · Ahmedabad Area, India

• Worked on developing and analysing the dynamics of network using community detection and other aspects of network theory of co-authors (authors publishing a research article). Data for this exercise was scraped from the public source Arxiv and the network analysis was done using the network visualisation tool Gephi.
• In another project I worked on the problem of modelling of dengue disease spread spatio-temporally using cellular automaton.