This repository is a centralized laboratory for Generative AI, Deep Learning, and Automated Data Governance. It contains end-to-end pipelines ranging from medical imaging classifiers to LangChain-based code analysis agents.
A specialized deep learning project focused on hematology analysis.
-
Dataset: Utilizes the BloodMNIST dataset (bloodmnist.npz).
-
Architecture: Features ResNet-based architectures defined in src/architecture.py.
-
Training: Interactive experiments documented in Jupyter Notebooks.
-
Inference: Production-ready scripts for testing against real-world blood smear samples.
Automating the understanding of software logic using Large Language Models (LLMs).
-
Agentic Workflows: Employs LangChain agents to analyze codebases.
-
Automation: Includes custom workflows for model listing and data ingestion.
Focuses on the ethical and secure handling of data in the AI era.
-
Governance Bot: An automated tool for enforcing data policy.
-
PII Masking: Demonstrates data privacy through masked customer datasets (customer_data_MASKED.csv).
-
Classification: Scripts to categorize data for compliance and security.
☁️ Cloud AI
Integration of AI services within cloud environments.
-
Architecture: Modular structure for deploying AI logic to cloud endpoints.
-
Privacy: Implementation of data masking for cloud-based processing.
A collection of utility scripts for the standard ML lifecycle.
-
Preprocessing: Robust data cleaning and preparation scripts.
-
Web Scraping: Automated tools for gathering training data from the web.
-
Deployment: Includes Dockerfile and app.py for containerized application hosting.
-
Deep Learning: PyTorch, Torchvision (ResNet), MedMNIST.
-
LLM Frameworks: LangChain, OpenAI/Open-Source Model APIs.
-
Data Science: Pandas, NumPy, Scikit-learn.
-
DevOps: Docker, Virtual Environments (venv).
-
Languages: Python 3.x.
It is recommended to use the provided venv structure:
# Activate the environment
source venv/bin/activate # Linux/macOS
.\venv\Scripts\activate # Windows
# Install dependencies
pip install -r requirements.txt
To ensure a consistent environment across all modules:
docker build -t gen-ai-hub .
docker run -it gen-ai-hub
- training_experiment.ipynb: The primary research document for the Medical CV project, containing loss curves, accuracy metrics, and model checkpoints (resnet_blood_best.pth).
This repository includes a Governance Bot and Data Classification logic to emphasize the importance of AI safety, PII protection, and responsible AI development. All customer data samples included are masked and dummy-generated for testing purposes.