• Create CI/CD pipeline to automate the process of model training, validation and model deployment.• Extract text from 1M+ PPTs by using the YOLO object detection model and achieved 87% F1-score.• Trained a transformer-based BERT text classification model, significantly improving accuracy by 23%.• Collaborated with cross-functional teams to design and develop a high-performance API for real-timeinference, leveraging AWS infrastructure and best practices for scalability and… Show more • Create CI/CD pipeline to automate the process of model training, validation and model deployment.• Extract text from 1M+ PPTs by using the YOLO object detection model and achieved 87% F1-score.• Trained a transformer-based BERT text classification model, significantly improving accuracy by 23%.• Collaborated with cross-functional teams to design and develop a high-performance API for real-timeinference, leveraging AWS infrastructure and best practices for scalability and reliability.• Created a scalable Kubernetes-based Kubeflow pipeline for end-to-end machine learning workflows,enabling seamless model development, training, and deployment in a production environment.• Utilise Docker and Kubernetes to containerise, orchestrate and scale over 12+ flask RESTful APIs.• Enhanced the performance of over 14+ machine learning models through rigorous hyper-parametertuning, resulting in an impressive 8% improvement in efficiency for detecting fraudulent claims. Show less

Implemented MVT architecture for developing the web application with the help of Django framework.Developed 60+ high-quality, high-performanced, maintainable APIs in django REST API framework for e-learning application.Designed frontend with react js & integrated with web API. Constructed Store Procedure in MS-SQL Server & optimized time complexity for 1M users.Extracted 2 Lakh+ rows of text data from images by performed OCR with tesseract as per the bussiness requirement.