• Ask me about Node.js, FastAPI, REST APIs, MongoDB, Docker, JWT Auth, ML, DL, NLP, LLMs, RAG Pipelines
• How to reach me: anandanaidupeyala@gmail.com
• B.Tech CSE @ RGUKT RK Valley | CGPA: 8.8 / 10.0
• Letter of Recommendation from NeuroStack Engineering Leadership
• Built Multimodal RAG pipelines combining textual retrieval with image-aware context processing
• Fun fact: I think backend magic is more exciting than frontend sparkles!

• Built and deployed ML inference APIs with FastAPI for low-latency, stable endpoints
• Designed data pipelines from model outputs to a Next.js frontend for consistent delivery
• Secured endpoints with JWT and RBAC; owned backend inference stability
• Supported end-to-end deployment of a production AI app on the Google Play Store
• 🏆 Awarded Letter of Recommendation for delivery, ML integration depth, and ownership
• Tech: Python FastAPI Docker JWT Linux Git

Real-time symptom analysis and risk triage powered by BART-large on Hugging Face Transformers

• Clean inference layer: preprocess → tokenize → predict → 0–100 risk score
• Triage engine maps risk to 4 care levels: self-care → consult → hospital → emergency
• Dockerized and deployed on Hugging Face Spaces + Render; tuned for low latency
• Tech: Python FastAPI HF Transformers (BART-large) Docker Hugging Face Spaces Render

Multi-role backend platform for community waste management with eco-rewards

• Developed using OOP and modular architecture in Node.js/Express.js
• Three-role system (Citizens / Collectors / Admins) with JWT-secured, role-specific API layers
• Coin-based eco-reward engine using atomic MongoDB transactions for data consistency
• Applied scalability design and performance optimization for production deployment on Render
• Tech: Node.js Express.js MongoDB JWT REST API Render

Explainable retinal screening with Grad-CAM visuals and instant PDF reports

• Fine-tuned EfficientNetV2-S (5 classes); 81.67% validation accuracy
• Pipeline: upload → preprocess (224×224) → predict → Grad-CAM → PDF
• FastAPI service on Hugging Face Spaces; frontend on Render; secure JWT/RBAC
• Tech: Python TensorFlow 2.15 Keras EfficientNetV2 Grad-CAM FastAPI Docker Hugging Face Spaces MongoDB Atlas

Retrieval-Augmented Generation system combining textual and image-aware context processing

• Built Multimodal RAG pipelines integrating textual retrieval with image-aware context processing for richer, more grounded responses
• Worked with multimodal embeddings, vector databases (FAISS), and retrieval orchestration across heterogeneous data sources
• Tackled real-world challenges of cross-modal alignment, retrieval relevance, and context fusion
• Focused on building AI systems that reason over both visual and textual information while remaining interpretable and reliable
• Tech: Python LangChain LangGraph FAISS Hugging Face Multimodal Embeddings Vector Databases

"Building scalable systems, one API at a time"