keyvanhamdi

Reduced internal report team workload by 70% by developing an automated web-based reporting system using LangChain + Flask. Optimized fraud detection rule sets to meet updated regulatory standards.

Increased data accessibility and decision-making quality by building an intelligent LLM-driven reporting service connected to the company Data Warehouse. Achieving first place in developing full anti-fraud and anti-money laundering compliance with regulatory laws between PSP companies in Iran. Developed a machine learning pipeline leveraging a local LLM model and LangChain, enabling enhanced visitor support for emergency maintenance tasks, reducing model implementation time by 50%. Designed and deployed ML models improving predictive accuracy across business workflows. Mentored junior data scientists in Python, SQL, and ML best practices.

Achieved 80% fraud case reduction through supervised and unsupervised ML methods. Increased POS deployment in high-transaction shops by 7%, improving market penetration. Identified nearly 200 profitable customers per month with a churn prediction model. Reduced emergency maintenance roll tickets by 75% using predictive modeling. Improved ETL pipeline performance by 25% via PySpark and SQL optimization. Supported segmentation strategies leading to onboarding 5,000+ new merchants. Built and enforced analytical KPIs using SSAS Multidimensional models for Power BI dashboards.

Increased online revenue by $50K/month through a new website and SEO strategy. Built automated crawlers for competitive pricing analysis, doubling pricing team efficiency. Managed cross-functional teams contributing to over $1.5M in brand revenue.