Building tech for justice systems in the global south to end judicial delays and improve access to justice

Helping a highly motivated team build the future of Customer Service using Autonomous AI Agents.LLM based QA

Worked as a Product Manager and Lead Engineer to launch the Zero to One Product called Equal Face. Which is a horizon-3 play for Equal. We discovered a niche problem, built a world class one of a kind solution and took it to production in 8 Months.Product:1. Discovered problem statements in Biometric Authentication based Identity for Indian Businesses. Built the first FaceID like Authentication experience for Android Devices as a Service.2. Built a team of 6 engineers and designers to execute on the Vision for Equal Face3. Did Successful POCs with EMRI, Indigo, RMZ and Aparna Constructions. These are leading buisnesses in India in Ambulance Service, Constructions and Aviation.4. Launched the product PAN India to leading businesses like EMRI.ML Engineering:1. Built our first face matching algorithm that’s at-par with the performance of largest vendor (Hyderverge) in context of our first customer2. Built out first version of passive liveliness model.3. Devled into high performance model quantisation for deployment to Edge Devices, both pre and post training.Backend/ML Ops Setup:1. Designed and wrote a backend alone and from scratch to handle millions of face matching transactions everyday. Modelled it on transactional DBs.Spent time on designing systems to cater to highest level of security and compliance of biometric data. Lots of great learning on best practices for infrastructure and deployment.

Product Engineering1. Built the ingestion and insight generation engine for Audio IntegrationsMachine Learning Engineering & Backend Development (June 2021 - Aug 2022):1. Built scalable ML engineering systems and contributed to development of the backend of Enterpret, bridging the gap between ML and product while ensuring smooth integration and resilience.2. Designed and implemented an ML platform for seamless inferencing of multiple models at a scale of over a million records per day and catering to the needs of ML researchers for training experiments.3. Gained proficiency in Python and Go programming language and deepened understanding of micro-services architecture in backend development.Machine Learning Researcher (Oct 2020 - May 2021):1. Built several maintainable Transformer Models for Sequence Labelling, Semantic Matching and Classification tasks and Language Generation.2. Successfully onboarded top-tier PLG companies like Notion, leading to funding from two world-renowned VCs.

1. Million scale Sentiment analysis and Aspect-based sentiment analysis.2. ML Infrastructure for Inference and Continuous Training

Intern at the Srini Raju Centre for Information Technology & Network Economy, Researching Features to Quantify Growth at Startups

Youngest intern ever at ITC infotech, joined right after my first year of college. Built a tabular data extraction model for internal invoices at ITC.