B.Tech CSE Undergrad at RCCIIT (Class of '27), passionate about architecting enterprise-grade Agentic AI systems and distributed full-stack applications.
My expertise sits at the intersection of robust software engineering, scalable cloud infrastructure, and applied machine learning.
🔭 Focus: Agentic AI (LangGraph) & End-to-End MLOps
🌱 Learning: Distributed Systems, Real-Time WebRTC & Ethical Hacking
💼 Experience: Architected EPIC, an AI-assisted Govt. Grievance Platform at Webel
🌐 Portfolio: anshunandi.vercel.app
🎨 Hobbies: Graphic Design, Video Editing & Competitive Programming
🏆 Awards: 2nd @ Cipher 7.0 | 3rd @ Codathon (Innovision '24)
Machine Learning · Deep Learning · NLP · Generative AI (LLMs, RAG) · Agentic AI
PyTorch · TensorFlow · LangChain · LangGraph · MediaPipe · Scikit-learn · Pandas · NumPy · Matplotlib · EDA
TypeScript · JavaScript · HTML/CSS · Tailwind CSS · React.js · Next.js · Express.js · FastAPI · Flask · NGINX · Streamlit · WebRTC · Socket.IO · JWT
AWS (S3, EC2) · GCP · Linux · Bash · Docker · Kubernetes · Git · GitHub · GitHub Actions · CI/CD · Celery · PostgreSQL · MongoDB · MySQL · SQLite · Redis · FAISS · Pinecone · ChromaDB · Supabase
Python · Java · C++ · C · TypeScript · JavaScript · SQL · Data Structures & Algorithms · OOP · DBMS · OS · Computer Networks
|
Architected a distributed full-stack Govt. grievance platform. Built a hybrid 4-tier LLM triage system, LangGraph ReAct agent, and FAISS RAG pipeline. Engineered a dual-provider email sync engine and real-time SSE dashboard. Enforced enterprise security with RLS, RISC webhooks, and a VRAM mutex. |
Architected a custom PyTorch CNN for binary chest X-ray classification. Engineered a modular MLOps pipeline (data ingestion → augmentation → training → evaluation) and a FastAPI inference server with automated CI/CD via GitHub Actions to AWS EC2 self-hosted runners. |
|
Engineered a Git-like CLI (yargs) with immutable SHA hashing and AWS S3 SDK sync for version-controlled blobs. Built a REST API with JWT/bcrypt auth and a GitHub-like React dashboard featuring contribution heatmaps and Socket.IO real-time collaboration. |
Engineered a P2P video conferencing app using WebRTC (ICE/SDP) and a Socket.IO signaling server. Implemented screen sharing, meeting rooms, call history tracking, and token-based session management with bcrypt hashing. |
|
Built a real-time CV pipeline (MediaPipe + OpenCV) computing 3D joint angles to render a live biomechanical HUD across 6 exercises. Integrated the Groq API for ultra-fast, context-aware LLM feedback and async gTTS audio cues. |
Engineered a multi-modal biometric system using 128-D facial embeddings (dlib + SVM) and librosa MFCC audio verification. Built a Supabase (PostgreSQL) backend with encrypted profiles, dual-role portals, and QR-code class-joining. |