🎓 MS Computer Science @ Western Michigan University
🚀 Building real-world AI systems, data tools, and trustworthy ML pipelines
- Build AI-powered systems (RAG, computer vision, ML pipelines)
- Design explainable and reliable ML frameworks
- Develop production-style Python tools (CLI + backend systems)
- Focus on real-world deployment and system design
- Retrieval-Augmented Generation (RAG) for reliable decision systems
- Explainable AI (XAI) in healthcare applications
- Data engineering + ML pipelines
- CLI tools & backend systems
- Evidence-grounded legal Q&A system
- Reduces hallucination via retrieval + validation
- Built with vector search and LLM integration
🔗 https://github.com/sum1tbarua/RAG-Based-Housing-Law-QA-System
- YOLOv11 for wound segmentation + ResNet-50 for anatomical classification
- Grad-CAM heatmap generation
- Mistral AI via Ollama for first-aid generation
- LLM-XAI using natural language explanation for interpretation
- Real-world healthcare AI pipeline
🔗 https://github.com/sum1tbarua/wound_detection_app
- Takes behavioral, demographic, and anthropometric details as inputs
- Performs real-time analysis and prediction of 3 classes (Diabetic, Prediabetic, and Non-diabetic)
- Uses explainable AI, SHAP, to break down feature contributions to prediction results
🔗 https://github.com/sum1tbarua/behavioral-diabetes-predictor
I focus on building:
- Trustworthy AI systems
- Explainable ML pipelines
- Real-world deployable AI applications
With applications in:
- Healthcare AI
- Secure AI systems
- Reliable LLM-based systems
