Abhinav Dutt

Working as a Data Scientist/Machine Learning Engineer in the Data Science Team providing solutions for Fraud Detection and New User Onboarding. Design and implementation of end-to-end pipeline for ML backed Collusion Detection Algorithms which reduced Customer Complaints related to cheating by 95%. Created ML classification model for detecting financial fraud using pyspark and XGBoost with a recall of 80%, reducing cyber cell complaints by 90%. Created ML models for detecting probability of New user playing cash games with precision of 95% and deployed with real-time latency using FastAPI. Identified Rings of fraudsters using Graph Databases which led to blocking of 1 million accounts.

Worked with the AI Team at Hero Electronix responsible for building ML/Deep Learning solutions for QUBO Home Security Camera and QUBO Baby Camera. Developed Image Classification Models with Pytorch for Baby Crying Detection, Mask Detection, Car Parking Detection. Developed APIs for sending recording and firebase event notifications for Person Detection, Motion Detection and Baby Cry Detection. Responsible for Design, Integration and Deployment of AI/ML features.