Headline

AI/ML Engineer | Fraud Detection · Credit Risk · GenAI | Data Analyst @ Onestop.AI (Ex) | Open to entry-level ML / Data Science roles

About

I built and deployed a full-stack AI credit scoring system — alone, as an undergrad — that predicts loan approvals at ~85% accuracy and explains every decision using SHAP. I'm a final-year B.Tech AI/ML student at NIMS University, Jaipur, building end-to-end ML systems at the intersection of financial risk, fraud detection, and explainable AI. My work spans the full pipeline: data wrangling on 300,000+ record datasets, feature engineering, model training with XGBoost, and production deployment on Render and Vercel. At Onestop.AI, I built NLP classifiers for hate speech and spam detection that crossed 90% precision — and reduced model error by 15% through systematic feature engineering and hyperparameter tuning across 10,000+ records. Right now I'm deepening my work in GenAI, LangChain orchestration, and LLM-powered financial applications — and building ChhatisPaisa, a finance + AI education brand that documents real money decisions for Indian students. Machine Learning · XGBoost · SHAP · Fraud Detection · Credit Risk · NLP · LangChain · Hugging Face · FastAPI · Python · Feature Engineering · Model Explainability Open to entry-level Data Scientist, ML Engineer, or Fraud/Risk Analyst roles — feel free to connect.

Location

Amber

AI stack

claude code

claude

cursor

github copilot

hugging face

langchain

Work experience

Founder & Content Creator at RitsonRacks, Apr 2026 - Present · 6 mo, India, Founded RitsonRacks (formerly ChhatisPaisa) to teach personal finance and AI tools to college students — creating content on real money decisions,and AI-powered money management without jargon. Built and grew a cross-platform audience across Instagram (@ritsonracks), YouTube, and Substack newsletter, publicly documenting my own financial journey using actual numbers.

AI/ML Developer at AI/ML Developer (Self-Initiated Projects), Jan 2026 - Apr 2026 · 4 mos, Jaipur, Rajasthan, India, Built an XGBoost loan approval classifier achieving ~85% accuracy; integrated SHAP for transparent, explainable fintech risk decisions — deployed as a full-stack FastAPI + React app with JWT auth on Render and Vercel. Performed large-scale EDA on 307,511 Home Credit loan applications across 122 features, identifying default risk drivers through correlation analysis, behavioral flags, and strategic missing-value imputation. Developed a TF-IDF + SMOTE NLP pipeline on 24,000+ records for imbalanced hate speech classification; benchmarked multiple models and selected the best by F1-score. Live Demo: https://ai-credit-scoring-system-omega.vercel.app/

Data Analyst Intern at Onestop.AI, May 2023 - Mar 2024 · 11 mos, Jaipur, Rajasthan, India, Built ML classifiers for hate speech and spam detection using Scikit-learn pipelines with TF-IDF vectorisation, achieving 90%+ precision across both tasks. Performed EDA on 10,000+ records using Pandas, NumPy, Matplotlib, and Seaborn; reduced model error rate by 15% through systematic feature engineering and hyperparameter tuning. Delivered weekly analysis reports translating model outputs into actionable insights for internal stakeholders.

Education

Bachelor of Technology - BTech, Artificial Intelligence, Nims university

Public links

github: https://github.com/Riteshchohan