🧠 Hello, I'm Shankar Tamang

AI/NLP developer passionate about turning research into real-world systems. I specialize in face recognition, chatbot frameworks, and building healthcare assistants that support professionals in their day-to-day tasks.

---

💼 Featured Projects

🧬 Liveness Detection for Face Recognition

Enhances facial recognition security with real-time liveness detection to prevent spoofing using printed photos or videos.

• 🔹 CNN models like LeNet, AlexNet, ResNet
• 🔹 Texture analysis and feature extraction
• 🔹 Spoof dataset augmentation and validation
• 🔹 Designed for integration into attendance systems

🔗 GitHub Repo: Liveness Detection Module

---

💬 LangChain-Powered Simple Chatbot

Lightweight conversational AI built using LangChain and Google Generative AI. Handles multi-intent classification and structured Q&A with document context.

• 🔹 RetrievalQA with Google GenAI embeddings
• 🔹 Chat memory and fallback responses
• 🔹 Structured intent detection with custom JSON
• 🔹 FastAPI + optional Streamlit frontend

🔗 GitHub Repo: LangChain Chatbot

---

🏥 Health Assistant for Medical Professionals

A streamlined medical assistant for doctors that generates structured patient reports, suggests medication classes, and allows PDF export of summaries.

• 🔹 LangChain prompt chaining and memory
• 🔹 Case documentation interface via Streamlit
• 🔹 Downloadable PDF reports
• 🔹 Ideal for use in clinics and training environments

🔗 GitHub Repo: Medical AI Assistant

---

🛠️ Skills & Tools

• AI/NLP: LangChain, Face Recognition, CNNs, LLMs
• Backend: FastAPI, Django, Flask
• Frontend: React, Tailwind CSS, Streamlit
• Utilities: OpenCV, MLflow, Conda, Git

---

🧭 My Mission

To build practical AI tools that are not just impressive — but genuinely useful. I’m driven to apply LLMs and CV to challenges in education, healthcare, and security, especially for Nepal.

Let’s collaborate on meaningful AI projects. 🌍