Machine Learning/AI Engineer

ramblr

AI for the Physical World.

At Ramblr, we build vision AI agents that understand work as it happens. Grounded in a company’s knowledge, processes, and operational context, our agents recognize real-world situations from egocentric video and other signals, guide operators through complex tasks, verify quality and compliance, and support service and maintenance teams at the point of work.

Ramblr One, our developer platform for physical AI agents, is the foundation for all of this: a data engine that structures video, images, manuals, and expert knowledge into activities, relations, and skills, plus an agent studio to build, evaluate, deploy, and monitor agents on our or the customer's infrastructure, with no model lock-in. Our customers range from industrial leaders to robotics teams closing the gap between data and physical AI.

Excited to join us at the intersection of AI and the physical world? Apply directly to this posting or use our open application form.

We look forward to hearing from you!

Job Description

We are looking for a strong Machine Learning/AI Engineer to deploy and maintain state-of-the-art models and embed them in production systems that combine vision-language models, computer vision, and agentic components. You write clean, efficient, and highly parallelized code to improve and extend our physical AI stack, and you help make building and deploying new features on Ramblr One faster and more reliable.

Drawing on your understanding of modern deep neural networks and agentic systems, you take responsibility for the performance and reliability of our platform. You build and optimize parts of our AI stack and integrate them efficiently into our products.

Your profile

Passion for solving the hard problems in physical and visual AI.

Required:

• Full proficiency in Python-based collaborative software engineering: consistent style guide, clean design patterns, self-documenting code, unit and integration tests, type annotations
• Vision-language and multimodal foundation models: production use beyond prompt engineering, including fine-tuning, evaluation, and integration into multi-component agentic systems
• Machine learning: data loading pipelines, serving and deployment, relevant neural network architectures, model optimization, PyTorch
• Scientific Python libraries such as NumPy and OpenCV
• General: Git, code reviews, development on Linux, distributed computing concepts
• A good mix of strong fundamental coding skills and effective use of modern coding assistants

Optional:

• Computer vision: Spatio-temporal data, image and video embeddings, augmentation techniques