Ankur Gupta

Engineered a high-performing Fraud Detection for fantasy sport achieving a remarkable 97.4% recall using hybrid algo encompassing Random Forest, KNN, Decision Tree, MLP and LR, leading to annual savings of up to 4 crores. Name Matching Algo for matching Pan name with Bank Name to identifying 10% more fraud. Defined the Skill Matching algo on real time for 2 player game bases of logistic function. Work on Image Detection using yolo for extracting card information from Playing Cards in rummy. Collaborate closely with Customer Retention team to defining strategies and analysis of campaign and funnels for fantasy and game.

Evaluating the performance of product features rolled out via A/B Testing experiment with google optimizer 360. Performing Inventory Planning and Store Replenishment for retail for pan India in EOSS time. Successfully developing a Delivery Model using the XGBoost to predict delivery probability with 80% of accuracy which impact to save 1 crore annually.

Architected, planned, designed, and implemented robust Incremental Data Warehouse on AWS EC2 instance. Developed DB procedures, resolved performance issues, & optimized SQL queries effectively. As a Developer, I’ve hands-on experience working on Odoo(openERP) on various modules, including Sales, Purchase, Manufacturing, Warehouse, Invoicing, POS, and Accounting.

Gathered images on basis of design/pattern to identify current trend in fashion jewelry by unsupervised learning by extract feature from transfer learning followed by PCA, K-mean. Collaborate with marketing team to monitor and optimize Facebook spending through Python, resulting in a remarkable 20% increase in daily Facebook Revenue. Achieved 20% higher Store Replenishment efficiency by employing RF for store prioritization.