AI Engineer - Connectomics Pipelines
Howard Hughes Medical Institute
Summary
You will develop and deploy production-ready pipelines for segmenting and proofreading connectomes from expansion microscopy imaging across fly, zebrafish, and Danionella. You will scale reconstruction from individual datasets to hundreds or thousands of connectomes, working with large-scale volumetric data on a compute cluster with around 800 GPUs.
Job description
About the role
You will develop and deploy production-ready pipelines for segmenting and proofreading connectomes from expansion microscopy imaging across fly, zebrafish, and Danionella. You will scale reconstruction from individual datasets to hundreds or thousands of connectomes, working with large-scale volumetric data on a compute cluster with around 800 GPUs.
What you will actually do
- Reconstruct connectomes at scale using pre-trained segmentation models
- Train and fine-tune models as new ground-truth data becomes available
- Design and run rigorous model evaluations to measure and improve reconstruction quality
- Optimize inference pipelines to increase reconstruction throughput on GPU clusters
- Build and support proofreading workflows (e.g., integration with CAVE)
- Coordinate manual annotation and proofreading efforts to improve training and evaluation data
- Release pipelines, models, and reconstructions to the broader scientific community with a focus on reproducibility
- Advise stakeholders on data strategy and best practices, and potentially mentor junior engineers
What we are looking for
- Trained and evaluated deep learning models (Transformers, U-Nets) for three or more years, ideally on 3D image data
- Deployed models with workflow orchestration and high-throughput processing on GPU clusters using tools such as Kubernetes, Nextflow, or LSF
- First-principles thinking: can explain why results move and design experiments to test explanations
- Strong Python in PyTorch or JAX
- Familiarity with modern vision models such as DINOv3, SAM, Cellpose, and Vision Transformers
- Agree on a measure of reconstruction quality first, then judge changes against it
- Curious about neuroscience and cell biology and enjoy working across disciplines
- Comfortable in a shared codebase with git-based collaboration
Nice to have
- Experience with microscopy image analysis, especially 3D electron or expansion microscopy
- Knowledge of connectome reconstruction methods such as Flood-Filling Networks (FFN), Local Shape Descriptors (LSDs), and graph agglomeration
- Experience with connectomics and microscopy tools such as CAVE, Neuroglancer, and Zarr
- Experience with self-supervised methods such as DINO or masked autoencoders
- Experience working with large biological datasets
Requirements
- Master's degree or PhD, or equivalent combination of education and experience
- Three or more years of deep learning experience
- Trained and evaluated deep learning models (Transformers, U-Nets), ideally on 3D image data
- Deployed models with workflow orchestration and high-throughput processing on GPU clusters using Kubernetes, Nextflow, or LSF
- Strong Python programming in PyTorch or JAX
- Familiarity with modern vision models such as DINOv3, SAM, Cellpose, and Vision Transformers
- Comfortable in a shared codebase with git-based collaboration
Benefits
- Competitive compensation package
- Comprehensive health and welfare benefits
- Retirement plans
- Time off
- Recognition and wellness programs
- On-site childcare
- Free gyms
- On-campus housing
- Shuttle bus service
Company overview
The Howard Hughes Medical Institute (HHMI) is a premier American non-profit biomedical research organization and philanthropy founded in 1953 by business magnate Howard Hughes. Headquartered in Chevy Chase, Maryland with additional facilities at the Janelia Research Campus in Ashburn, Virginia, HHMI is one of the largest private funding organizations for biological and medical research in the United States. With an endowment of approximately $24 billion, it ranks as the second-wealthiest philanthropic organization in the country. HHMI supports over 300 investigators at more than 60 host institutions across the nation, investing about $825 million annually in biomedical research. The institute focuses on fundamental biological research across diverse fields including genetics, immunology, molecular biology, neuroscience, and structural biology. Beyond research funding, HHMI also supports innovative science education programs, curriculum development, and public engagement initiatives to inspire the next generation of scientists. The organization operates with approximately 2,300 employees and maintains a commitment to open science, requiring immediate open access to publications from its funded research since 2022.