Software Engineer/Founder

Depika Gupta  • Peerlist

Introducing 'ChatPDF' (Deployed on Streamlit) - a revolutionary solution for accessing insights from PDFs, especially those with current information. Unlike other LLMs, 'ChatPDF' excels in accurately addressing questions based on document content. The methodology involves multiple PDF loading, text chunking, language model utilization for vector representations, similarity matching, and response generation, ensuring reliable and informed decision-making. By leveraging the capabilities of Python, LangChain, HuggingFace's LLM, Streamlit, HTML, CSS, Git, and GitHub, this project ushers in a new era of information extraction from PDFs, addressing the limitations of pre-trained models.

ChatPDF - OpenAI's GPT-3.5 Integrated Chatbot

An AI-powered medical assistant web app designed to provide personalised healthcare recommendations, connect users with specialists, and offer mental health support.

NeuroLift

PresenterX is a Computer Vision and NLP intergrated Django web-app that represents a transformative approach to presentation delivery through gesture and voice-command control, empowers presenters with an immersive toolset. It allows presenters to seamlessly navigate slides, highlight content, and engage with the audience using intuitive, hands-free gestures and voice instructions, making presentations more dynamic and interactive, enhancing the presentation experience for both presenters and audience.

PresenterX – Gesture and Voice-Activated Pres

In a digital landscape overflowing with information, deciphering key insights from lengthy blog posts is a persistent challenge. 'SummarEase' addresses this issue by utilizing state-of-the-art Hugging Face NLP Transformers, specifically BART, I evaluated various pre-trained models such as GPT-2, BERT, and BART, identifying BART as the top performer based on the ROUGE scores (ROUGE-2 and ROUGE-L) under consideration in Blog Post Summarization. The GPT and BERT models are among the most popular extractive summarization models and the BART model is the abstractive summarization models. And, hence, I leveraged the BART model using Deep Learning techniques and implemented it as a Flask web application, 'SummarEase' that simplifies information retrieval, offers rapid, succinct summaries, enhancing content access and reducing energy consumption by 16%.