muhammad abdullah

I'm currently an AI Developer intern and Data Science undergraduate at Virtual University, building Python-based machine learning applications with API integrations. My work includes Aegis Earth AI, Veritas AI, and Veritas Guard—AI tools for climate analysis, misinformation detection, and fraud verification.

I built Veritas AI with a multi-stage verification pipeline that extracts claims through LLM reasoning, cross-checks credible news sources via APIs, and returns confidence-scored verdicts with citations. For Veritas Guard, I applied similar reasoning and web search verification to identify scam red flags and assess fraud risk.

I've also developed churn and credit-card fraud models, achieving 84.8% ROC-AUC for churn prediction and 80% PR-AUC with 72.6% recall for fraud detection. I'm seeking to grow as an ML/Data Science Engineer through applied, real-world problem solving.