Ajay Gurav

Pioneered AI training program that employees contribute in AI related use cases and tasks. Redesigned existing early warning fraud detection system with ML to a 40% increase in Accuracy. Brainstormed new use cases for Trade Product and led team to deliver those.

Built and Deployed Time Series Forecasting Model to predict order volume at each Last Mile Hub. Improved accuracy by 11% by using the Middle out approach. Managed and led the Data Science team and held end-to-end responsibility for delivery, planning, conceptualizing, developing, deploying & tracking data science solutions. Developed a labor forecasting model to predict ideal labor at each Facility hub and achieved 86% accuracy and reduction in manual labor planning by 28%. Implemented Simulation Dashboard to adjust forecast based on additional insight. Automated Invoice Details Data Extraction from Scanned and Non-Scanned Invoices that reduced Manual Labor by 40%. Architected End to End Data Engineering and Model Training Pipeline with Vertex AI (MLOPS GCP framework).

Collected, studied, and interpreted large healthcare data and implemented data pipelines to reduce the data processing time by 30%. Co-Founded Doconvid.AI and raised $100K in funding with the Nvidia Inception program. Mentored Junior Data Scientists in projects from a technology standpoint. Led the implementation of an end-to-end machine learning pipeline that saved computational cost for training by 70%. Created the report, performed accurate and successful data management, and showcased model prototypes with 85% base accuracy. Facilitated meaningful improvement to existing NLP models through carefully directed research, resulting in a 15 % increase in accuracy. Designed and implemented algorithm for asthma risk stratification and performed EDA, feature engineering, outperformed statistical models by 20 % in term of accuracy.

Analyzed old information architecture and advised on the design and development of the new one. Conducted training sessions on the software developed and presented to the management for the approval of deployment to reduce manual workload by 50%. Advised Information, feedback, and guidance to clients to support technology related decision making.

Designed an Automatic Voucher Match system to cut down manual work by 80%. Architected Dashboard app for real-time drug health monitoring system with the IoT Orchestration service. Collaborated with industry professionals to generate innovative ideas to improve the automation process resulting in a 23 % increase in revenue.