Signature Verification Service:-- R&D of model using Image Processing and CNN-based image classifierChat Assistant for Query Resolution:-- Contributed in the development of Chat Assistant using OpenAI Engine - Developed Kubernetes based Rest Services (Java Spring & Python Flask) to serve user request- Utilised AWS based MySQL DB for data persistanceEmpowered Search experience at Upstox- Phase-2 -- Researched over leading Search Engine Distributors like AWS OpenSearch, ElasticSearch & MongoDB Atlas Search -- Experience with AWS OpenSearch APIs (Data ingestion, searching, etc)- Phase-1 -- Developed scalable backend Search Service serving Multi-Million request on daily basis -- Contributed in framing the System Design for End-to-End Stocks Search Service -- Contributed in the service development as both service owner & individual contributor with features such as typo-tolerance, dynamic sorting, trending search, personalisation, etc. -- Service handling Millions of request per day -- Utilised various tech involving TypeSense search, AWS services, various NLP techniques for search, Python, JavaQ&A Service using LLM- Contributed in the High-Level System Design- Contributed in development of the service by utilising various tech involving OpenAI, AWS, NLP and vector-searchOther Contributions in Data Sciences & Analytics:- - Various Analytical strategies to support marketing campaigns (User-groups, Data Insights, Proposal) - Forecasting use-cases using various regression and time-series approaches (Monthly Active Users, User Propensity Modelling)- POCs on use-cases related to Object-Detection & Image Recognition- Experience designing complex SQL queries - Utilising various AWS resources (Lambda, Athena, Sagemaker, Studio, Pipelines, Cloudwatch, S3)- Experience in building Rest APIs using Python flask & Java Spring Boot

Innovation Team mavQ (Startup backed by MTX)- Applying SOTA machine learning and deep learning algorithms to different modalities of data such as tabular data, documents, text, and images to solve real-life problems.- Competitor analysis and market research on the targeted use case.- Extraction of information from the text in the form of named entities and Personal identifiable information utilising different NN models.- Information extraction from documents such as key-value pairs, named entities utilising different utilising LayoutLM, Yolo, etc.- Model training & hyper-parameter tuning.- Data preparation: cleaning, augmentation, annotation- Developing python-flask APIs and deployment in the cloud to server the solutions.- Developing end-to-end custom training pipelines.