Senior VLM & VFM Field Engineer, Matroid

• A master’s degree or higher in computer science, computer engineering, electrical engineering, machine learning, artificial intelligence, robotics, or a closely related technical discipline
• Strong experience with visual recognition tasks such as detection, segmentation, tracking, action recognition, and anomaly detection, plus hands-on work adapting and evaluating visual foundation models or vision-language models
• Excellent Python skills and sound software engineering practices for building systems that connect models, APIs, databases, and customer infrastructure
• Proficiency with relevant tools and frameworks, such as PyTorch, TensorFlow, JAX, Hugging Face, NumPy, and OpenCV
• Experience developing rigorous evaluations and improving AI performance through better data, model adaptation, prompts, architecture, or workflow design
• Demonstrated ability to make architecture decisions and resolve reliability, performance, and integration problems in deployed systems
• Experience leading complex technical projects and collaborating across teams
• Clear written, verbal, and presentation skills in English, including the ability to explain technical decisions and tradeoffs to both technical and business audiences

Desirable

• We’re hiring a Senior Vision-Language Model (VLM) & Visual Foundation Model (VFM) Field Engineer to build and deploy production-grade deep learning systems using Matroid’s platform for a broad range of enterprise customers, such as Mercedes-Benz, Toyota, Boeing, Google, Amazon, Bosch, and many others
• You’ll work directly with customers to understand their operations and turn challenging requirements into reliable computer vision and multimodal AI solutions. As the technical owner of complex engagements, you’ll guide architecture, model adaptation, evaluation, integration, and deployment, combining Matroid’s platform with custom engineering
• Projects will cover defect inspection, anomaly detection, assembly verification, process and safety monitoring, and multimodal understanding, retrieval, and reasoning across images, videos, and operational data. Alongside delivery, you’ll use lessons from customer deployments to improve our platform and engineering practices
• Build and deploy deep learning systems on Matroid using visual foundation models, vision-language models, and/or specialized vision-centric models for object detection, segmentation, tracking, anomaly detection, spatiotemporal reasoning, and multimodal understanding and retrieval
• Own technical delivery from scoping and feasibility assessment through model selection, fine-tuning, prompt engineering, system architecture, and production deployment
• Develop deep learning systems that can operate in cloud, edge, and hybrid environments, optimizing inference latency, GPU utilization, memory consumption, and throughput while balancing accuracy, cost, and maintainability
• Establish representative evaluation datasets and task-specific benchmarks. Analyze failure modes and improve model accuracy and robustness through data curation, targeted augmentation, and DL system adaptation
• Develop imaging and data collection systems, selecting cameras, optics, illumination, and capture configurations that support reliable visual recognition in demanding environments
• Connect Matroid to customer equipment and software, including PLCs, manufacturing execution systems, SCADA, quality management platforms, safety alerting, and video management systems
• Guide customers through technical decisions and troubleshoot complex issues across model behavior, inference pipelines, infrastructure, and integrations
• Review architecture and code, and build reusable tools for model evaluation, inference, and deployment
• Work with product and engineering colleagues to translate deployment insights into improvements in model capabilities, performance, and platform workflows
• Deliver technical workshops, training, and documentation that help customers evaluate, adopt, and operate their AI systems successfully
• Take ownership. Bring structure to ambiguous requirements, identify risks early, and drive projects through production readiness
• Stay hands-on. Write code, adapt models, investigate data, and debug systems while providing technical direction
• Choose models thoughtfully. Assess capabilities and limitations, and select the right combination of fine-tuning, prompting, specialized models, and orchestration for each application
• Build deep platform expertise. Understand Matroid’s capabilities and apply its features, workflows, and integrations effectively
• Understand customer operations. Account for production constraints and communicate clearly with engineers, operators, and business stakeholders
• Raise the team’s technical standard. Share deployment lessons, offer practical feedback, and help other engineers develop their judgment
• Work independently and collaboratively. Make progress with autonomy and bring in teammates when their expertise will improve the outcome
• Keep learning. Evaluate advances in computer vision and multimodal AI and turn promising approaches into practical customer capabilities