Data Scientist
Tehran, Iran
• Had meetings with potential clients across a variety of sectors• Determined potential customers' data maturity and developed a roadmap for improvement• Assessed and prepared proposals for potential customers' problems (e.g., Estimation of Origin-Destination Matrices Using Traffic Counts, Pricing Mechanism for Premium SIM Cards, Calculating Health-Index for Water Pumps, Tire Tread Depth Estimation)Providing an ML solution for a utility company:• Ingested data and performed… Show more • Had meetings with potential clients across a variety of sectors• Determined potential customers' data maturity and developed a roadmap for improvement• Assessed and prepared proposals for potential customers' problems (e.g., Estimation of Origin-Destination Matrices Using Traffic Counts, Pricing Mechanism for Premium SIM Cards, Calculating Health-Index for Water Pumps, Tire Tread Depth Estimation)Providing an ML solution for a utility company:• Ingested data and performed EDA on 124 tables (10M+ records) of the billing system• Built a data-centric learning pipeline to increase data quality• Designed a data transformation method to calculate overlapping time intervals, lowering processing time by 70%• Implemented various time series clustering and forecasting methods, identifying previously unseen consumption patterns• Implemented a novel ensemble clustering method to detect optimal clusters automatically, reducing the workload for developers significantly• Implemented a Robust Random Cut Forest (RRCF) model for customer consumption patterns, improving anomaly detection by 29% and fraud detection by 13% Show less