Aman Sharma

About Aman Sharma

Connected Vehicles Data ScientistI work on: vehicle safety (active and passive), applied statistics, deep learning, and computer vision.

Experience

1. Data Scientist

   Ford Motor Company

   Aug 2021 — Present · Chandler, AZ, US

   Driver Score Model for Fleet Customers: Built a mathematical penalty-based driver scoring model to reduce idling and warranty costs.Driver Behavior Optimization with Genetic Algorithm and RL Agent: Designed a multi-objective genetic algorithm with a reinforcement learning agent to recommend effort-efficient behavior changes for improving driver scores.Stolen Vehicle Detection Model Using Siamese Neural Networks: Developing a deep learning model using Siamese Neural Networks (SNNs) to detect high-risk vehicle theft patterns based on behavioral deviation.Wheel Vibration Monitoring Using Harmonics and Generative Models: Developing an unsupervised anomaly detection system leveraging harmonic energy signatures and autoencoders to identify wheel imbalance issues.OTA DTC Spike Detection via Sequential Sample Ratio Mismatch: Applied sequential statistical testing (SSRM) to detect post-OTA spikes in diagnostic trouble codes, improving issue detection speed and reliability.Distracted Driving Alert System: Led the analytics pilot of distraction behaviors using multi-modal sensor data and causal analysis, to quantify distraction rates through optimized alert timing.Vehicle Classification Using GMM Clustering: Built a semi-supervised Gaussian Mixture Model to segment vehicle usage patterns and uncover high-risk driver clusters.

Education

• Texas A&M University

  Master of Science - MS, Transportation- Autonomous Vehicles

  2017 — 2019

• Panjab University

  Bachelor of Engineering (B.E.), Civil Engineering

  2012 — 2016

• Udacity

  Intro to Self-Driving Cars Nanodegree Program, Robotics

  2020