Kowsar Asghari

– Architected AI-powered resource allocation framework using Multi-Agent RL and Graph Neural Networks processing 42,807 NASA satellite detections; achieved 38.7% improvement in response time and 23.4pp success rate increase validated through discrete-event simulation – Designed and deployed geospatial analytics pipeline integrating terrain data, weather APIs, and spatial network optimization across 31 provinces; implemented conformal prediction achieving 94.7% empirical coverage for uncertainty quantification – Conducted benchmark validation against commercial emergency dispatch systems; performed scalability analysis for multi-region deployment with cost-benefit modeling and causal inference for counterfactual policy evaluation

– Developed demand forecasting models for urban air mobility market improving forecast accuracy 25%; built competitive intelligence dashboards in Looker Studio analyzing 30+ market demand indicators for strategic planning – Designed Python and PySpark validation pipelines coordinating QA processes across 150+ markets ensuring data quality standards; managed nationwide CRM infrastructure and data monitoring systems – Collaborated with product and operations teams delivering predictive analytics for market expansion; presented data-driven recommendations to executive stakeholders supporting $2M+ investment decisions

– Engineered supply-demand matching algorithms optimizing provider scheduling across healthcare services; structured semantic data layers and automated BI reporting reducing manual reporting hours 85% – Implemented demand forecasting pipelines using PySpark for chemical supply chain reducing stockouts 22% and improving operational efficiency $120K annually; built validation scripts ensuring business logic compliance – Designed data models for production scheduling and inventory management; coordinated with Marketing, Operations, and Management teams delivering analytics solutions supporting 5+ business units

– Developed predictive customer scoring models using classification algorithms (Random Forest, XGBoost); designed and executed 50+ A/B experiments with statistical significance testing (t-tests, chi-square) improving targeting precision 30% – Built customer segmentation pipeline using K-Means and hierarchical clustering; automated churn analytics dashboards in Tableau tracking 100K+ customer interactions reducing churn 18% – Owned NPS analytics program analyzing 10,000+ B2B survey responses using dimensional analysis achieving +7.8% improvement; defined KPI frameworks for CRM teams coordinating analytics impacting 200K+ users

– Developed airbase location optimization model for wildfire suppression using K-Means clustering on 20 years of NASA MODIS satellite fire data; applied stochastic optimization for resource allocation under uncertainty – Built geospatial data processing pipeline integrating satellite detections, terrain analysis, and distance calculations; research recognized as KANS Competition Finalist (Mustafa Foundation, Transportation Field)