2026 PhD Residency, Computer Vision (Tapestry)

This six-month PhD residency focuses on pioneering the application of Vision-Language Models (VLMs) to automate the understanding of critical electrical infrastructure and defect detection. By bridging the gap between raw imagery and actionable insights, you will help Tapestry build a more visible and resilient global power grid.

How you will make 10X Impact

• Lead the curation and high-quality labeling of specialized datasets focusing on electrical assets and complex infrastructure defects.
• Design and execute rigorous evaluation frameworks to benchmark Vision-Language Model performance on domain-specific imagery.
• Partner with research scientists to fine-tune state-of-the-art models, improving their ability to reason about physical infrastructure.
• Analyze and visualize model performance bottlenecks to drive iterative improvements in image understanding.
• Support the exploration of model distillation techniques to bring advanced research capabilities toward real-world deployment.
• Collaborate across engineering and research teams to integrate vision insights into the core Tapestry digital twin.

What you should have:

• Currently pursuing a PhD or MS in Computer Science, Electrical Engineering, or a related technical field with a focus on machine learning.
• Strong foundational knowledge in deep learning, particularly within the domains of computer vision and natural language processing.
• Proficiency in Python and experience with deep learning frameworks such as PyTorch, JAX, or TensorFlow.
• Practical experience managing, preprocessing, and analyzing large-scale image datasets for research projects.
• Ability to communicate complex technical concepts clearly to both specialized researchers and general engineering partners.
• A curious, research-first mindset with a drive to see theoretical models solve tangible, physical-world problems.

It’d be great if you also had one or more of these:

• Direct experience working with Vision-Language Models or multi-modal architectures.
• Exposure to large-scale distributed training environments and data pipelines.
• A portfolio of peer-reviewed publications, significant open-source contributions, or innovative research projects.
• A creative approach to problem-solving when faced with sparse or noisy real-world data.