SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Cyvl is a Physical AI company building infrastructure intelligence for cities and towns across the US. The company develops purpose-built sensors, computer vision, and AI systems that turn vehicle-collected data into actionable intelligence about road conditions, assets, and infrastructure health. Over 500 cities use Cyvl's platform, and approximately 1 in 30 Americans drive on roads managed with Cyvl data.
As an AI Engineer, you will work with one of the largest proprietary datasets of US infrastructure—LiDAR and imagery from hundreds of cities—to build AI systems that detect, classify, and assess infrastructure assets. Your work will span computer vision models for pavement, signs, sidewalks, and trees; LLM-powered products including agentic workflows and AI copilots; and model deployment and monitoring in production.
Key responsibilities include:
- Building and improving computer vision and 3D perception models that detect and assess infrastructure from LiDAR and imagery
- Developing LLM-powered products: agentic workflows, MCP servers, and AI copilots that integrate Cyvl data into customer tools
- Owning models end-to-end: from data and training through evaluation, deployment, and production monitoring
- Designing evaluation sets and metrics that align model performance with customer needs
- Staying current with state-of-the-art AI/ML techniques and integrating relevant tools into products
The role is based in Somerville, MA, with a hybrid model (on-site ~90% of the time). Remote work from San Francisco is available with periodic travel to Boston for key milestones.
Requirements:
- 3+ years building and shipping ML or AI systems in production
- Strong Python and modern deep learning framework experience (PyTorch preferred)
- Hands-on experience with computer vision, 3D perception, or LLM applications (RAG, tool use, agents, evals)
- Solid software engineering fundamentals: testing, code review, cloud deployment
- Ability to move from ambiguous problems to working prototypes quickly, then productionize
Nice to have:
- Point cloud or multimodal work (LiDAR and imagery fusion)
- Model Context Protocol (MCP), foundation model APIs, or agent frameworks
- Geospatial or remote sensing background
- MLOps experience: model serving, GPU infrastructure, experiment tracking
- Advanced degree in CS, ML, robotics, or related field (or equivalent experience)
