Contact
HomeProjectsAboutBlog Contact
๐Ÿ”

RAG Knowledge Assistant

A production-grade Retrieval-Augmented Generation system built with LangChain, FAISS, and HuggingFace Sentence Transformers. Ingests multi-format documents (PDF, DOCX, web), chunks intelligently, embeds to a local vector store, and answers questions with source citations. Supports swappable LLM backends (Ollama, OpenAI, Groq).

LangChainFAISSHuggingFaceOllamaFastAPIPython
๐Ÿ‘๏ธ

Diabetic Retinopathy Detection

Deep learning pipeline for automated diabetic retinopathy severity grading from retinal fundus images. Uses a fine-tuned EfficientNet backbone with custom augmentation strategy for handling class imbalance in medical imaging. Achieves clinician-grade accuracy on the APTOS benchmark.

PyTorchEfficientNetMedical ImagingOpenCVPython
๐Ÿ“‹

Insurance Claims Processing (GCP)

End-to-end ML pipeline for automated insurance claims classification and routing on Google Cloud Platform. Combines Document AI for structured extraction, BERT-based NLP for intent classification, and BigQuery ML for risk scoring. Deployed on GKE with KubeFlow orchestration and real-time monitoring.

GCPDocument AIVertex AIBigQuery MLKubeFlowGKE
Model Card Advanced RAG System v2: Evaluation Summary
Faithfulness
0.92
Answer Relevancy
0.88
Context Precision
0.85
Research Time Saved
40%