Eon Systems
Senior Segmentation ML Lead - Immediate Start
Eon Systems, San Francisco, California, United States, 94199
About us:
At Eon, we are at the forefront of large-scale neuroscientific connectomic data collection. Our mission is to enable the safe and scalable development of brain emulation technology to empower humanity over the next decade, starting with a digital twin of a mouse.
Role We are seeking a Senior Segmentation ML Lead to spearhead the segmentation phase of our connectomics project. This individual will be responsible for leading and managing the efforts aimed at transforming high-resolution brain images into a complete connectome. This work will span the entire pipeline, from distortion correction, normalization, 3D registration, segmentation, to database structure, and plausibly automated proofreading. You will spearhead this work on your own for a few months as you and we begin to hire a team of engineers for you to manage.
Responsibilities
Segmentation Algorithm Optimization and Deployment:
Lead the optimization and deployment of existing segmentation algorithms on expansion microscopy datasets, including flood-filling networks (FFNs) and local shape descriptors (LSDs)
Novel Algorithm Development:
Lead the development and implementation of a novel segmentation algorithm with efficiency in mind, likely based on a Swin transformer architecture and using distillation techniques
Preprocessing Workflow Management:
Use open-source or develop new algorithms for an efficient preprocessing workflow at high data rate, including multi-view deconvolution
Cross-Departmental Collaboration:
Collaborate with data engineers and hardware teams to ensure optimized deployment of preprocessing steps and segmentation across thousands of non-collocated GPUs in a highly distributed cloud
Team Leadership and Scaling:
Manage a team of computational neuroscientists and ML engineers to scale and improve our segmentation algorithms
Automated Proofreading Solutions:
Work with automated proofreading solutions like RoboEM to further reduce segmentation errors without manual intervention
Possibly:
create our own automated SOTA proofreading method utilizing multiscale inputs (might be housed on separate team)
Skills
7+ years of experience
in machine learning, deep learning, computer vision, or image segmentation, ideally with an emphasis on EM segmentation
Demonstrated ability to manage and optimize large-scale segmentation projects (or related equivalent)
Excellent understanding of image processing and normalization techniques
Experience managing and mentoring junior team members and providing technical leadership in machine learning practices
Preferred Skills:
Familiarity or implementation experience with FFNs, LSDs, and other embedding models
Familiarity with specific large-scale connectome image processing techniques like CLAHE, CycleGAN, lens distortion correction, and 3D registration
Experience with distributed systems, cloud, and high-performance computing
What We Offer:
Competitive salary and equity compensation
Opportunity to work on a burgeoning field that impacts neuroscience and AI safety
A chance to take ownership of a critical aspect of our connectomic project, with significant computing resources and the most cutting-edge tools
Management experience
A fast-paced work environment with many opportunities for growth
#J-18808-Ljbffr
Role We are seeking a Senior Segmentation ML Lead to spearhead the segmentation phase of our connectomics project. This individual will be responsible for leading and managing the efforts aimed at transforming high-resolution brain images into a complete connectome. This work will span the entire pipeline, from distortion correction, normalization, 3D registration, segmentation, to database structure, and plausibly automated proofreading. You will spearhead this work on your own for a few months as you and we begin to hire a team of engineers for you to manage.
Responsibilities
Segmentation Algorithm Optimization and Deployment:
Lead the optimization and deployment of existing segmentation algorithms on expansion microscopy datasets, including flood-filling networks (FFNs) and local shape descriptors (LSDs)
Novel Algorithm Development:
Lead the development and implementation of a novel segmentation algorithm with efficiency in mind, likely based on a Swin transformer architecture and using distillation techniques
Preprocessing Workflow Management:
Use open-source or develop new algorithms for an efficient preprocessing workflow at high data rate, including multi-view deconvolution
Cross-Departmental Collaboration:
Collaborate with data engineers and hardware teams to ensure optimized deployment of preprocessing steps and segmentation across thousands of non-collocated GPUs in a highly distributed cloud
Team Leadership and Scaling:
Manage a team of computational neuroscientists and ML engineers to scale and improve our segmentation algorithms
Automated Proofreading Solutions:
Work with automated proofreading solutions like RoboEM to further reduce segmentation errors without manual intervention
Possibly:
create our own automated SOTA proofreading method utilizing multiscale inputs (might be housed on separate team)
Skills
7+ years of experience
in machine learning, deep learning, computer vision, or image segmentation, ideally with an emphasis on EM segmentation
Demonstrated ability to manage and optimize large-scale segmentation projects (or related equivalent)
Excellent understanding of image processing and normalization techniques
Experience managing and mentoring junior team members and providing technical leadership in machine learning practices
Preferred Skills:
Familiarity or implementation experience with FFNs, LSDs, and other embedding models
Familiarity with specific large-scale connectome image processing techniques like CLAHE, CycleGAN, lens distortion correction, and 3D registration
Experience with distributed systems, cloud, and high-performance computing
What We Offer:
Competitive salary and equity compensation
Opportunity to work on a burgeoning field that impacts neuroscience and AI safety
A chance to take ownership of a critical aspect of our connectomic project, with significant computing resources and the most cutting-edge tools
Management experience
A fast-paced work environment with many opportunities for growth
#J-18808-Ljbffr