Computer Vision Engineer

Role Summary

As a Computer Vision Applied Research Scientist, you will work at the intersection of cutting-edge research and production AI systems, developing multimodal vision models capable of understanding complex construction drawings. You'll own the full research lifecycle—from designing experiments and developing novel architectures to deploying production-ready models that directly impact customers.

Responsibilities

• Design, train, and evaluate state-of-the-art computer vision and multimodal AI models.
• Develop solutions for segmentation, detection, OCR, and document understanding tasks.
• Conduct rigorous experimentation and translate research findings into production systems.
• Build and optimize training pipelines using modern vision transformer architectures.
• Deploy and maintain production-ready models in real-world inference environments.
• Collaborate closely with leadership, including the CTO and Head of Engineering, on technical direction and research priorities.
• Drive improvements in model accuracy, robustness, and scalability.
• Contribute to the team's research culture through experimentation, technical discussions, and knowledge sharing.

Desired Skills

• 3–7+ years of professional experience in Computer Vision, Applied AI, or Machine Learning.
• Advanced English level.
• Strong hands-on experience with modern vision transformer architectures (SAM, DINO, CLIP, SigLIP, or similar).
• Strong experience with PyTorch and deep learning model development.
• Experience with segmentation, detection, OCR, document understanding, or multimodal vision-language systems.
• Experience taking models from research environments into production.
• Strong understanding of deep learning fundamentals, including optimization, regularization, loss functions, and self-supervised learning.
• Ability to work during California business hours.

Plus Experience

• Graph Neural Networks (GNNs) or relational reasoning architectures.
• OCR, document understanding, or layout analysis.
• LoRA and parameter-efficient fine-tuning techniques.
• Self-supervised pretraining methods (MAE, DINO, etc.).
• Engineering drawings, construction documents, or technical diagrams.
• Published research at top-tier conferences (CVPR, ICCV, ECCV, NeurIPS, ICLR).
• Open-source contributions in Computer Vision or Machine Learning.
• Production inference optimization, deployment, and model serving.

Who You Are

• You thrive in autonomous environments and are comfortable with ambiguity.
• You take ownership and enjoy driving research and experimentation end-to-end.
• You are curious, proactive, and focused on solving challenging real-world problems.
• You communicate clearly, collaborate well, and can support technical decisions with evidence.