⚖️ AI Fairness Training Gym — Detect, Measure, Fix & Explain bias in AI models using RL (PPO) + Gemini AI | Google Solution Challenge 2026

Explainable AI-powered telecom fraud detection system using Random Forest, Isolation Forest, Rule-Based Intelligence, SHAP Explainability, FastAPI, and Streamlit Dashboard for real-time fraud risk assessment.

AI-powered customer churn prediction system with explainable AI (SHAP), interactive dashboard, and conversational insights using LLMs.

Decision Copilot is a CLI-first, LLM-powered decision analysis tool that runs structured multi-agent reasoning pipelines with full persistence, explainability, and reproducibility.

RAD-XAI is an explainable-AI framework for chest X‑ray pneumonia detection that benchmarks CNNs and ViTs using Grad‑CAM-based saliency maps and quantitative XAI metrics (concentration, faithfulness, agreement) to audit model reasoning and surface clinically unsafe failure modes beyond standard AUC/accuracy.

Explainable AI system for CV evaluation, field detection, and personalized feedback

ML-powered customer churn prediction with SHAP explainability — TheCalebGlobalAnalytics (TCG)

Stima dell'età cerebrale da risonanze MRI 3D: confronto tra Machine Learning classico e CNN 3D, addestrati interamente su CPU. Include interfaccia Streamlit con interpretabilità Grad-CAM.

AI-powered IPL analytics platform with match prediction, team/player insights, and Explainable AI (XAI) — built with Flask, scikit-learn & Plotly.

🛡️ AI-powered tender evaluation system for CRPF procurement. Reduces 30-45 days of manual bid evaluation to under 2 minutes using Explainable AI, OCR, NLP & Blockchain Audit Trails.

Explainable AI dashboard for German Credit Risk Prediction using XGBoost and SHAP.

Heart Failure detection model using Explainable AI (Decision Trees).

AI-powered Telecom Customer Churn Prediction System using XGBoost, SHAP Explainability, and Streamlit Dashboard.

CNN-based image classifier with xAI explainability techniques — Saliency Maps, Grad-CAM, LIME, RISE and uncertainty rejection. Built with PyTorch and Optuna.