Machine Learning Engineer IV, Data

ACV Auctions is hiring a Machine Learning Engineer IV to build end-to-end computer vision solutions for its vehicle inspection platform. The role focuses on developing damage detection models, operating high-throughput model-serving infrastructure, and improving inspection speed, valuation accuracy, and system reliability at scale.

Responsibilities:

• Design and train high-performance computer vision models for automated damage detection, focusing on precision, recall, and model robustness
• Architect and maintain high-throughput, containerized microservices for model serving using REST/gRPC to ensure low-latency performance
• Collaborate with business stakeholders to translate complex inspection requirements into scalable, production-grade ML solutions
• Own the end-to-end model lifecycle, from experimentation and design to deployment and optimization in high-traffic environments
• Design and maintain robust data pipelines using Kafka to ensure high-fidelity inputs for model serving and inference

Requirements:

• Graduate education (MS or PhD) in a computationally intensive domain or equivalent work experience
• 5+ years of prior computer vision experience
• Advanced proficiency with Computer Vision frameworks (e.g., PyTorch, OpenCV, TensorFlow) and Python/SQL
• Experience designing and maintaining visual data annotation pipelines and evaluation frameworks for complex, real-world image datasets
• Experience optimizing high-latency models for real-time inference
• Backend software engineering experience in the cloud (AWS / GCP) with a focus on microservices (docker) and the ML model development lifecycle
• Experience building and maintaining streaming data pipelines (e.g., Kafka) for real-time model serving
• No immigration or work visa sponsorship provided for this position
• Knowledge of ML frameworks and libraries, such as Kubeflow, Databricks, KServe and so on
• Experience designing evaluation frameworks for complex visual data
• Experience leading technical design reviews