Projects

FileGenie: Your Documents, Instantly Answered

Python, RAG, Langchain, AWS, OpenAI, Flask, React, LLM

  • Engineered a full-stack RAG application enabling users to upload PDFs and get contextual answers using Langchain and FAISS vector database for efficient document retrieval and embedding generation.
  • Implemented session-based user isolation for secure multi-user support and deployed on Render with isolated storage for each user session, achieving fast and accurate document-based responses.
Live Test

Product Advertiser : Transform Products Into Stunning Advertisements

Hugging Face, Stable-Diffusion, OpenCV, Python

  • Developed a model using Hugging Face, Stable Diffusion, and OpenCV to generate advertisement visuals from product images guided by prompts.
  • Applied computer vision techniques for background manipulation and enhanced image quality, ensuring brand alignment and visual appeal.
Live Test

Vessel Identification in Dialysis Treatment

Python, YOLO(v7), LabelImg

  • Collected a dataset of vein and artery images, enhancing YOLO v7 model training accuracy by 20% and reducing false positives by 15%.
  • Used labelImg for data annotation and applied mAP metrics to evaluate classification accuracy, collaborating with medical experts for refinement.

Subtheme Sentiment Analysis

Python, Spacy, NLTK, Text Preprocessing

  • Developed a sentiment analysis pipeline using Spacy and NLTK to extract and analyze subthemes from customer reviews.
  • Implemented text preprocessing and pronoun resolution techniques to ensure accurate sentiment attribution for each subtheme.
Live Test