Research Engineer -- Vision and Vision-Language Models

What you'll do

• Develop and evaluate vision-language models through pretraining, post-training, reinforcement learning, and large-scale training runs.
• Source, generate, curate, process, and evaluate vision and multimodal datasets.
• Improve model architectures, training methods, and training and data-processing performance for visual-agent tasks.

What they ask for

• Strong machine-learning intuition and ability to take ambiguous problems from conception through experimentation and execution.
• Strong implementation and prototyping ability, with capacity to learn unfamiliar technical areas quickly.
• Proficiency with PyTorch and Python.
• Ability to collaborate and communicate effectively across research and engineering.

Nice to have

• Experience with multimodal models or datasets, large-scale training, reinforcement learning, vision datasets, or image-processing pipelines.
• Experience with computer-use or visual agents, frontend or mobile development, large-scale GPU clusters, or published machine-learning research.

Tools

PyTorch, Python

AI focus

Research and engineering for vision-language models, multimodal training, reinforcement learning, and agentic visual tasks.