Projects

Customer Satisfaction

Developed an end-to-end customer satisfaction prediction project utilizing ZenML for pipeline management, MLflow for model tracking, and Streamlit for production deployment. Achieved successful model deployment with real-time predictions, demonstrating expertise in machine learning project development and deployment.

StreamlitMLOpsZenmlMLflowMachine Learning

Fine Tuning Mistral-7B-v0.1 model for Research Paper Summarization

Fine tunned Mistral-7B-v0.1 model hugging face open source model on custom dataset of research paper. Also used prompting techniques to show result of final model after fine tuning.

Fine TuningPre-trainedLLMPythonNLPSummarizationGPU

Class Conditional Image Generation

Designed and implemented class-conditional image generation models tailored for alphanumeric and animal datasets within a 24-hour hackathon. These models were trained on proprietary data and deployed on a Django web server environment. Through user input prompts, the system generates corresponding images, showcasing a seamless integration of deep learning and web development technologies (Django, Python)

Diffusion ModelTransformerAttentionGPUPythonTensorflowPytorch

My Website

This website, crafted(using template) and coded by me, is a dynamic platform where I consistently enrich it with cutting-edge resources and insights on Artificial Intelligence, aiming to provide visitors with an ever-growing repository of knowledge and innovation.

PortfolioWesiteWeb Development

ICapNet 2023

The proposed method for automatic image captioning combines LSTM, CNN, and Attention Mechanism based on Flickr8k Dataset. It uses CNN to extract image features, LSTM to generate word sequences, and attention mechanism to focus on relevant image parts during captioning. This approach aims to improve caption accuracy and relevance.

LSTMCNNAttentionGPUPythonTensorflow

Xpecto-2022

Developed the first-ever techfest website ”Xpecto” for a college, using React-based frontend technology.Utilized expertise in React, HTML, CSS, and other web technologies to create an engaging and interactive website.Played a key role in the successful launch of the website, receiving positive feedback from users

ReactHTMLCSSResponsive designProblem-solving

Ace Dynamics

The Credit Card Fraud Detection Problem involves using Logistic Regression to model past credit card transactions,including fraud and non-fraud cases, to identify new transactions as fraudulent or not. Accuracy was measured using various techniques, and the model was deployed using Flask framework for production use.

PythonHTMLCSSFlaskJavascriptML-AlgorithmProblem-Solving

Machine Learning in Fluid Mechanics

A Machine learning approach in fluid Mechanics for finding Velocity Field and Pressure field. This project helps to understand that how machine learning is expanding in each domain of science. In this project via ideas of Deep learning that is artificial neural network (ANN) how we can predict different flow patterns of fluid.

Deep LearningFluid Mechanics

Kaggle Expert

I participate in various competitions on Kaggle and utilize Kaggle notebooks along with their GPU for multiple research projects and competitions. You can view all the code I've written there.

Problem SolvingPythonDLML Algorithm

Hand Gesture using Computer Vision

This is an Open-CV project build by me using CVZone library. This project basically helps for presenting ideas through PPT or image and teacher can hightlight points just using his fingers from any distance.

PythonOpenCVComputer Vision

Flipr-Hackathon

Platform made using php for truck trivers for managing there workload

phphackathon

AI Thesis Evaluation Tool

The code provided utilizes the OpenAI language model to evaluate the quality of a thesis based on predefined evaluation criteria.

Fine TuningPre-trainedLLMPythonNLPSummarizationGPU