This repo is developed and maintained for education purpose to learn Agentic AI by superml.dev
We introduce our first project with Agentic AI, available to learn Fraud Detection in Card data Authorization.
Available at https://github.com/superMLdev/smart-sdlc Smart-SDLC uses role-aware AI execution, where each persona operates within explicit delivery boundaries, owns specific artifacts, and hands work forward through governed transitions.
Available at https://github.com/crazyaiml/nl-2-sql-agent A guarded NL-to-SQL reasoning agent that learns schema and domain context, then answers analytical questions using iterative, evidence-backed SQL.
Available at https://github.com/superMLdev/anti-fraud Its a public repo for learning purpose only.
This project is a modular fraud detection platform that uses:
- 🧠 Traditional Machine Learning models (XGBOOST)
- 🤖 LLMs (Large Language Models) for human-style explanations (OPEN AI can be extended to other LLMs)
- 📈 Drift detection to retrain models when fraud patterns change
- 🔁 Reinforcement Learning-style feedback loops
- Real-time fraud prediction based on transaction data
- Transparent LLM-based explanations of decisions
- Feedback loop for manual corrections and re-training
- Modular pipeline for drift detection and re-training
- API endpoints for inserting transactions and getting predictions
Built with ❤️ by crazyaiml @ SuperML.dev