Full Resume
Pankaj Kumar
Data Scientist · ML, GenAI & Agentic AI
Professional Summary
Data Scientist with 1+ year of experience in machine learning, GenAI, and agentic workflows. Expertise in building and deploying scalable ML systems and RAG pipelines using FastAPI and cloud platforms.
Professional Experience
Data Scientist
June 2025 - PresentTransOrg Analytics, Gurugram, Haryana, India
- Analyze and transform large, multi-source datasets to support machine learning and business decision-making.
- Design and develop end-to-end ML solutions, including feature engineering, model training, evaluation, and deployment.
- Optimize pricing and revenue strategies using predictive modeling and data-driven experimentation.
- Build and maintain AI/GenAI applications, including RAG systems and agentic workflows for enterprise use cases.
- Develop scalable data governance and quality-monitoring workflows to improve data reliability and reduce manual effort.
- Create executive dashboards and insight tools that translate complex data into actionable business outcomes.
Key Projects
Seat Selection Revenue Optimization
- Built an ML pipeline (Logistic Regression, LightGBM, XGBoost) on a 40M+ record airline dataset to predict paid seat-selection propensity.
- Designed a predict-then-optimize pricing engine to find revenue-maximizing SSR prices, lifting paid seat-selection rate and ancillary revenue.
Agentic Data Governance
- Architected a multi-agent governance platform (FastAPI, AutoGen, LangChain, LangGraph) with 13+ AI agents, cutting manual governance effort by ~75%.
- Built automated data lineage, an 8-dimension quality assessment pipeline, and a RAG chatbot with dual retrieval (Qdrant + Neo4j) and RBAC.
Financial Intelligence RAG Platform
- Built a REST API-based RAG assistant using hybrid retrieval (dense + BM25) and reranking to turn financial reports into natural-language Q&A.
- Added LLM-as-judge evaluation and confidence scoring, tracked with LangSmith and MLflow, to monitor reliability in production.
Personal Projects
SquadSplit
squadsplit.pankajk.devSelf-hosted PWA for group expense management with real-time balance tracking and one-tap debt simplification, serving 100+ users.
FastAPI, SQLAlchemy, Azure VM, CI/CD
CruxAffairs
cruxaffairs.pankajk.devSelf-hosted AI-powered current affairs PWA with automated scraping, LLM summarization, and daily quiz generation for exam aspirants, serving 50+ active users.
FastAPI, Airflow, Docker, LangChain, Azure VM, CI/CD
Skills
Programming & Data
Python, SQL (PostgreSQL/MySQL)
AI & Machine Learning
GenAI (RAG, LLMs), NLP, Agentic Frameworks (LangGraph, AutoGen, LangChain), Classical ML (Regression, Classification, Clustering), Prompt Engineering
Backend & MLOps
Airflow, FastAPI, MLflow
Cloud & DevOps
Microsoft Azure, Docker, Git, CI/CD (GitHub Actions, Azure DevOps)
Education
B.Tech, Production & Industrial Engineering
Indian Institute of Technology Roorkee | 2021 - 2025
Completed coursework in optimization, operations management, statistics, and industrial systems.