My Work

$\"\"$

What I do

From understanding your requirements, designing a blueprint and delivering the final product, I do everything that falls in between these lines.

Fine-Tuning Large Language Models (GPT, LLaMA)

Fine-tuned LLMs for medical applications using PyTorch, Transformers, and CI/CD pipelines with GitHub Actions, ensuring reproducibility with MLflow.

Text Summarisation

Developed an NLP-based text summarisation solution using Azure Cloud, FastAPI, and MLOps practices with CI/CD, Docker, and MLflow for deployment and scalability.

RAG-Powered Financial Planning Assistant

Designed a Retrieval-Augmented Generation (RAG) system utilising FAISS, LangChain, and NLP for intelligent query responses, integrating LLM-based virtual assistant capabilities.

Autonomous Driving Car Detection

Developed and deployed an Object Detection system using the YOLO model, leveraging computer vision and imagery data to achieve accurate and efficient real-time car detection.

Face Recognition

Developed a Convolutional Neural Network (CNN) using TensorFlow and Keras, which accurately identified and verified individuals in various images.

Image Segmentation with U-Net

Developed a U-Net model for accurate image segmentation using deep learning and computer vision techniques.

Customer Churn Prediction Model

Developed a logistic regression model for customer risk prediction and designed an ETL pipeline for data preprocessing and business insights.

Credit Card Fraud Detection

Developed a Machine Learning model using logistic regression, leveraging EDA to address the class imbalance and improve fraud detection accuracy.

Predictive Modelling and Analysis for Drilling Data

Built predictive models and applied data pre-processing, visualisation, and cross-validation for accurate and generalised insights.

Art Generation with Neural Style Transfer

Implemented Neural Style Transfer with VGG19, leveraging CNNs and Transfer Learning to blend artistic styles with content images.

Image Classification Using Transfer Learning

Developed an image classifier using transfer learning with MobileNetV1, incorporating data augmentation, custom layers, and performance evaluation for accurate classification.

AI-Powered Jazz: Generating Solos with LSTM Networks

Developed an LSTM-based generative model for jazz solo improvisation, leveraging sequence modelling, deep learning, and MIDI data processing with TensorFlow and Music21.

RAG-Powered Medical Assistant

Developed a multimodal RAG app for medical document analysis using Generative AI, Langchain, and FAISS to enable intelligent retrieval, summarisation, and question-answering across text and images.

Real-Time Stock Data Fetcher

Developed a real-time stock data fetcher using Python, integrating web scraping, data manipulation, Dockerisation, and CI/CD with GitHub Actions for automated deployment and testing.

Generate Blogs using LLaMA

Developed a blog generation app using LLaMA, enabling customisable content creation with real-time responses, tone selection, and audience targeting.

Interactive PDF Chatbot with LangChain

Developed a chat application using LangChain and OpenAI GPT, enabling interactive PDF analysis with text extraction, vector storage in Cassandra, and intelligent question-answering.

Character-Level Text Generation Using GPT and Transformers

Developed a character-based text generation model using GPT and Transformer networks in PyTorch, applying deep learning techniques for creative NLP tasks with model training and deployment.

Transformer-based Named-Entity Recognition

Implemented a Transformer-based Named-Entity Recognition (NER) system using Hugging Face, fine-tuning pre-trained models to accurately classify entities in text.

Smart Emoji Recommender

Developed an emoji recommender model using deep learning, LSTM, and pre-trained GloVe embeddings to enhance text expressiveness with relevant emojis.

Speech Recognition – Trigger Word Detection

Developed a trigger word detection system using deep learning with LSTM-based RNNs, enabling real-time speech recognition and accurate audio event detection.

Car Price Prediction Model

Built a car price prediction model using Lasso and Linear Regression with data preprocessing, feature engineering, and regularisation for improved accuracy and generalisation.

Flower Classification Model

Built a classification model using SVM and KNN, applying feature engineering and model evaluation to ensure accurate species identification..

Adult Census Income Classification Model

Developed a machine learning model, focusing on data preprocessing, feature selection, and model evaluation for accurate income prediction.

Netflix Movie Recommendation System

Built a recommendation system using collaborative and content-based filtering, applying data preprocessing and similarity metrics for personalised suggestions.

Image Classification Model using CNNs

Built an image classification model using CNNs, leveraging deep learning for feature extraction, training, and accurate image recognition.

AI-Powered Chatbot Using Knowledge Graphs

Developed a knowledge chatbot using Neo4j, LangChain, and OpenAI, integrating a knowledge graph for environmental insights and LLM-powered recommendations.

Object Detection and Analysis Platform

Developed and deployed a computer vision platform integrating NLP to analyse images, detect specific objects, and provide actionable insights to enhance user interaction and decision-making.

Lung Cancer Detection Using Gradient Boosting

Developed a machine learning model for lung cancer detection, achieving high accuracy through advanced pre-processing, feature engineering, and optimisation while mitigating model bias in a research study.