Project 1: How AI Reads and Understands Mortgage Documents

I analyzed how AI interprets mortgage documents by studying Machine Learning, Deep Learning, and NLP techniques. I explored OCR and Computer Vision, working with a real mortgage 'blob' file to understand how AI handles both structured and unstructured data.

Project 3: Data Extraction from Documents Using Python

I converted both digital and scanned PDFs into machine-readable formats using Python. I leveraged tools like PyMuPDF and pdfplumber, applying heuristics to extract key fields such as borrower names, salaries, and loan amounts from multi-page mortgage documents.

Project 5: Introduction to Retrieval-Augmented Generation (RAG)

I learned to build a Retrieval-Augmented Generation (RAG) pipeline using LlamaIndex, which facilitates efficient data retrieval from large document sets. This project enhanced my understanding of how to optimize AI models for quick information access.

Project 7: Blob Processing, Classification, and Routing

I designed a system to process and classify unstructured mortgage file blobs into individual documents. By using layout-based clues and rule-based methods, I effectively routed documents to the appropriate extraction logic, streamlining the workflow.

Project 9: Final Integration, Testing, and Evaluation

I integrated all components into a complete document intelligence system, rigorously testing its performance on various document types. I packaged my work into a professional demo, showcasing the system's capabilities through a Colab repo, documentation, and walkthrough video.

Project 2: Learn to Work with Python for AI-Powered Document Processing

I utilized Python for data processing, focusing on cleaning and organizing mortgage documents to prepare them for AI automation. I gained hands-on experience with Google Colab, mastering data preparation techniques and enhancing OCR accuracy through image processing.

Project 4: Advanced OCR Comparison and Layout-Aware Extraction

I compared three OCR engines—Tesseract, PaddleOCR, and EasyOCR—on complex mortgage documents. I learned to clean and extract text from scanned PDFs and explored layout-aware tools to preserve formatting, ultimately recommending the best tools for real-world applications.

Project 6: Advanced RAG and Open-Source Experiments

I advanced my RAG pipeline by implementing chunking techniques to improve document segmentation and metadata-based filtering. I evaluated the performance of open-source models against proprietary options, optimizing the system for real-world mortgage documents.

Project 8: Gradio Chatbot with RAG Integration

I transformed my RAG pipeline into an interactive chatbot using Gradio. This project allowed users to interact with the system through a web-based UI, making it easier to query and retrieve information from PDF documents.

Project 1: Understand the Operational Challenge

I examined the operational challenges Amazon faces regarding employee retention, particularly focusing on the significant turnover within the first 90 days of employment.

Project 3: Find Sentiment & Keyword Patterns in Employee Voice Data

Project 2: Collect & Clean Employee Voice Data

I collected and cleaned employee voice data from various platforms to ensure it was ready for analysis, setting the stage for deeper insights into employee experiences.