AI Engineer with 2+ years of experience building production-grade AI systems, Agentic AI applications, Retrieval-Augmented Generation (RAG) pipelines, Computer Vision solutions, and data analytics platforms.

I specialize in designing intelligent systems using LLMs, LangGraph, LangChain, FastAPI, Computer Vision, and modern AI infrastructure.

• 🔭 Building Agentic AI and Multi-Agent Systems
• 🧠 Developing Enterprise RAG Applications
• 🤖 Creating LLM-powered products and workflows
• 📊 Experienced in Data Science, Analytics, and MLOps
• 🌱 Exploring GraphRAG, MCP, Fine-Tuning, and AI Agents
• ☁️ Working with AWS, Docker, FastAPI, and Cloud Deployments

Python • SQL • C++ • JavaScript

TensorFlow • Scikit-Learn • PyTorch • LangChain • LangGraph • OpenAI • Groq • Hugging Face • Fine-Tuning • LoRA

FAISS • ChromaDB • GraphRAG • Hybrid Search • Cross-Encoder Reranking

FastAPI • Flask • REST APIs • SSE Streaming

Snowflake • Power BI • Pandas • NumPy • ETL Pipelines

AWS EC2 • AWS S3 • AWS Lambda • AWS SageMaker • Docker • GitHub Actions

PostgreSQL • SQLite • Supabase

Built enterprise AI solutions including:

• Agentic Chat with Database using LangGraph
• Multimodal RAG Systems
• SQL Agents and Document Intelligence
• LLM Observability using LangSmith
• AML Risk Analytics Platform
• Knowledge Graph Solutions for Google Projects

✅ Resolved 77% of customer queries using Agentic AI

✅ Improved retrieval accuracy by 28% using Query Expansion and Cross-Encoder Reranking

✅ Reduced LLM operational cost by 20% using prompt and token optimization

✅ Processed 300,000+ AML assessments in 90 seconds

Full-stack AI interview platform featuring:

• GPT-4o-powered interview generation
• Deepgram speech-to-text
• OpenAI TTS voice responses
• Resume-aware questioning
• Real-time streaming responses
• Camera gaze monitoring using TensorFlow.js & MediaPipe
• Google OAuth via Supabase

Tech: React, FastAPI, OpenAI, Deepgram, Supabase, TensorFlow.js

AI fitness coach that analyzes exercise form in real time.

Features:

• Real-time pose estimation
• Squat, Push-Up, Curl, Shoulder Press & Lunge tracking
• AI voice coaching using Llama 3.3 70B
• Text-to-Speech feedback
• Workout history analytics

Tech: Streamlit, MediaPipe, OpenCV, Groq, SQLite

Multi-modal attendance system using:

• Face Recognition
• Voice Recognition
• QR-based enrollment
• Automated attendance marking

Tech: dlib, Resemblyzer, Scikit-Learn, Supabase, Streamlit

Multi-agent research platform that:

• Plans research strategies
• Performs web searches
• Synthesizes reports
• Sends email summaries
• Maintains research sessions

Tech: OpenAI Agents SDK, GPT-4o, Streamlit, SQLite

Medical research paper classification system.

• NLP pipeline using TensorFlow
• Deep Learning based sentence classification
• Hyperparameter tuning
• Achieved significantly higher accuracy than baseline ML models

Tech: TensorFlow, NLP, Deep Learning

• Microsoft Azure Data Scientist Associate
• Hugging Face – Fundamentals of Agents
• IBM – Python for Data Science, AI & Development

📧 Email: in.shubhamshekhar@gmail.com

💼 LinkedIn: https://linkedin.com/in/shubham-coder/

🐙 GitHub: https://github.com/ShubhamZoro

I enjoy building AI systems that combine LLMs, Computer Vision, and Agentic Workflows to solve real-world business problems.