• 🎓 B.Tech, Computer Science & Engineering - ITER, SOA University (2022 – 2026)
• 🛠️ Ex-Data Engineering Intern at MathCo, building Azure Databricks/PySpark ETL pipelines
• 🔭 Currently building multi-agent LLM systems (LangGraph, RAG, MCP) and real-time streaming data pipelines (Kafka, dbt, Airflow)
• 🌱 Currently deepening my skills in AI systems, Machine Learning, and Data Engineering
• 💬 Ask me about ETL architecture, agent orchestration, or why my pipeline chose 3 AM to fail on a Friday

aqi-pipeline - Real-time air quality ELT system Hourly pipeline ingesting live CPCB government AQI data across 10 Indian cities. Kafka (Redpanda) → Postgres → dbt (implements the official CPCB AQI methodology) → Airflow → Streamlit dashboard, fully containerized with Docker Compose. Python Kafka dbt Airflow PostgreSQL Docker Compose Streamlit

QuizForge - Multi-agent AI coaching platform A quiz app that doesn't stop at grading — a LangGraph agent diagnoses weak subtopics, researches and writes a lesson using RAG with cited sources, then re-quizzes you. MCP tool server for agent tool-calling, structured outputs, LLM-as-judge validation. TypeScript Next.js Python FastAPI LangGraph RAG pgvector MCP

Blood_Report_Analyser - OCR-based lab report ETL pipeline Ingests unstructured PDF/image blood reports, extracts and validates 26+ parameters against reference ranges, and flags abnormalities and cross-parameter correlations through a config-driven rules engine. Deployed on Streamlit Community Cloud. Python Pandas pdfplumber OCR Streamlit

Languages:

Data & Cloud:

AI & LLM:

Frameworks & Tools:

• 📍 Based in Bhubaneswar, Odisha, India
• 🎬 Self-appointed cinema connoisseur- my movie reviews are more rigorous than my code reviews
• 🎨 Graphic designer since 14- professionally opinionated about fonts, colors, and whitespace since before I could vote
• 🧪 Side projects run on real government data, real bugs, and real 2 AM commits
• 📈 If a pipeline runs clean on the first try, I don't trust it either