Ajitesh Bhan

Blog post Generator model: Instagram and blog post Gen AI model with complete automation, utilizing Jasper AI. ChatBot for travel website based to handle dynamic customer queries. Enhanced the chatbot with a Retrieval-Augmented Generation (RAG) framework that integrated vector databases ( FAISS) and OpenAI embeddings, enabling accurate, real-time retrieval of context-specific information from a large dataset. Automated the chatbot pipeline using AWS Lambda and containerized deployment with Docker, achieving a 40% improvement in response time. Developed and optimized scalable NLP models for analyzing extensive textual data from diverse sources, including social media, news outlets, and blogs. Led end-to-end workflows covering data collection, preprocessing, model training, hyperparameter tuning, and deployment, ensuring efficient and accurate insights from large datasets

Below are examples of few of my projects: Object Coordinate detection: Built computer vision model capable of determining the coordinates of the detected objects with Accuracy rate 88%. Image preprocessing. Code optimisation and factorisation. Technical stack: Opencv, Python, pytorch. Document management systems : Developed a Hugging face LLM model for Text Tagging. Document Analysis, data cleaning and formatting. Gathering of business requirements. Automation of reports including data and graphical representations. Technical stack: Python, Hugging Face API, pytorch, R, ggplot. Medical Image Screening : Improving the quality of medical images and preparing the data set (shot variant, image optimization and image enhancement by background subtraction). Classification of medical images into classes (tumor/malignancy classes, effective networks and Resnet experiments). Front end development using Fastapi Technical stack: Python, Opencv, fastapi, R, ggplot, tensorflow.

Development of Neural Network for People counting Project with accuracy of 90 percent. Utilizing TesorFlow Object Detection API for final Model Development. Deployment of Model to Production Pipeline (Quantization and Minimization of Loss on Quantization). Implemented 3D convolutional Network for detecting the behavior action of human and animal cases using Pytorch

Database Management using SQL. Building data Pipeline and Integration of datasets. Generating Reports and Charts (Data Visualization using Python and SQL Integration using tableu . Applying Statistical Analysis and Machine learning models for text analysis on customer review.