• Developed an asynchronous task execution with complex workflows using Celery, RabbitMQ & Django for AI based content researchplatform.• Engineered comments analysis tool to derive audience insights on Youtube videos using GPT-3.5 with efficient sampling andaggregation.• Developed an analysis report generation flow using RAG architecture and Chain-of-Thoughts prompting technique.• Conducted summarization of lengthy documents utilising techniques such as Map-Reduce within LangChain.• Implemented a conversational Q&A pipeline with RAG architecture using LangChain, Milvus, and MongoDB.• Created copilots using GenAI to provide real-time analytics insights with dynamic visualisations as responses.• Contributed to the development of an Image Processing AI pipeline tailored for detecting and cropping comic panels to improvemobile readability of comics. Additionally, crafted a JavaScript based Photoshop script to arrange panels sequentially, streamliningworkflow and effectiveness.• Technologies: Generative AI, Langchain, Celery, OOPs, MongoDB, Vector Database, Django, Google Cloud Platform.

• Devised robust vehicle routing system with Google OR-Tools and Machine Learning.• Implemented an advanced Face Recognition and Fraud Detection system utilising Artificial Intelligence, which improved order execution efficiency by 15% and minimised fraud incidents by 40%.• Designed and implemented an optical character recognition system to enhance user experience by verifying card expiry.• Developed a recommendation system that integrated Redis for efficient caching, resulting in a remarkable increase in website engagement time by more than 60%.• Technologies: FastAPI, Python, Google OR-Tools, AWS, OpenCV, OCR.

• Developed an AI chat bot for customer support with Rasa, resulting in an impressive 80% reduction in manual support time.• Created a recommendation system, using machine learning algorithms and MySQL, for suggesting relevant jobs to users.• Gained expertise in ML algorithms and acquired valuable skills in data cleaning and wrangling techniques.• Technologies: Python, Rasa, Flask, Machine Learning algorithms, MySQL.

• Conceptualized with professor at University of Calgary for research on federated learning.• Helped in making various defense techniques from scratch.• Created reports on various research papers related to federated learning.

• Generated annotations for various exercise videos.• Learnt about posenet models for detecting different human poses.• Helped in creating virtual private trainer for the user.

• Scraped data using selenium in python.• Trained models to predict the price of various commodities for different states of India.• Helped in training time-series models.