Member of Research Staff (Machine Learning for Neural Circuit Modeling Postdoc)

About Episteme

Episteme is a new kind of R&D company built for people who want their work to matter in the world. We support exceptional researchers pursuing ambitious, translational science; work that often struggles to find the right home within traditional academic or corporate institutions. At Episteme, we provide the financial, infrastructural and operational support that allows researchers to focus on the problems that genuinely deserve their attention.

We work closely with researchers to move ideas from early insight to real-world application. Translation is part of the work, not an afterthought. When breakthroughs are ready, we help bring them into the world responsibly, including through commercialization when appropriate.

Working at Episteme means joining a company that is intentionally different. We are building something new, and that requires comfort with ambiguity, intellectual honesty, and a bias toward thoughtful execution. You will work alongside people from many disciplines who care deeply about rigor, impact and follow-through. We value clarity over complexity, ownership over passivity and collaboration without micromanagement. Roles evolve through contribution and trust rather than rigid job descriptions, and ideas are tested quickly, decisions are made explicitly, and learning is shared openly so the system improves faster than any individual.

What We Look For

This is a Member of Research Staff role, equivalent in scope and career stage to a postdoctoral researcher position. The position is about developing statistical and machine learning models that process, explain, and predict neural dynamics from large-scale experimental recordings.

You will work within the research group of Dr. Michael Skuhersky, contributing to a broader effort to decode and model neural dynamics in C. elegans by combining experimental observations with computational models. Working closely with both computational and experimental scientists, you will develop and deploy machine learning models for imaging data, neural activity traces, and predictive and causal modeling of neural circuits.

You have strong foundations in machine learning and statistical modeling and are excited by the challenge of applying those tools to difficult scientific questions. You understand that biological systems are noisy, incomplete and often resistant to simple explanations. Rather than treating this complexity as a limitation, you see it as an opportunity to build better models and ask better questions.

You operate with a high degree of independence within a defined research scope. You can take broad objectives and translate them into experiments, analyses, and models while exercising sound scientific judgment. You know when to push forward independently and when to seek input from collaborators.

You are motivated by learning, discovery, and scientific rigor. You care not only whether a model performs well, but also whether it improves understanding, generates useful hypotheses, and helps move the field forward.

Research Focus Areas

• Predictive modeling of neural activity from optical recordings
• Integration of multimodal priors including imaging, behavior, and connectomics
• Statistical inference and causal analysis of circuit-level dynamics
• Development of advanced machine learning approaches including graph neural networks, symbolic regression, mechanistic models, and statistical learning methods
• Mathematical frameworks connecting learned representations with mechanistic hypotheses of neural and circuit function

Key Responsibilities

• Develop and test ML models for processing and integrating multi-modal data priors (e.g., calcium imaging, connectomics, and voltage dynamics of individual neurons and circuits)
• Develop and test ML models for predicting neural activity and behavior from experimental datasets
• Explore advanced deep learning methods (e.g., symbolic regression, GNNs, Transformers, generative models) to inform neural models from data and enhance interpretability
• Apply statistical and causal inference methods to identify candidate circuit mechanisms
• Develop models and software that can be broadly adopted by the scientific community through open-source or commercial distribution models.
• Contribute to the collaborative, interdisciplinary environment of the project
• Follow a structured research plan, with defined tasks and milestones.

What Success Looks Like

Success is reflected in the quality, rigor and usefulness of the work you produce.

Machine learning models developed by the group become more predictive, interpretable and scientifically meaningful because of your contributions. Experimental and computational researchers are able to learn faster because your work helps clarify which hypotheses deserve further investigation and which do not.

Your models generate genuine scientific insight rather than simply improving benchmark performance. They help connect observations, mechanisms, and predictions in ways that improve understanding of neural circuit function.

You consistently deliver high-quality research with minimal supervision, make sound decisions within your scope of responsibility, and contribute positively to the progress of the broader research program.

Even when experiments or models fail, they produce useful learning that improves future work.

Over time, your contributions become a trusted part of how the team understands neural dynamics and approaches increasingly difficult scientific questions.

Who Thrives in This Role

You have strong Machine Learning foundations and a deep interest in applied neuroscience.

You are technically strong but intellectually humble. You care about evidence, are willing to revise your assumptions, and are comfortable operating in areas where answers are not known in advance.

You understand what model features will increase the utility of the model to the relevant scientific and industrial communities.

You are excited by difficult, high-dimensional datasets and enjoy developing new approaches when existing methods fall short. You can move comfortably between mathematical reasoning, software implementation, model development, and scientific interpretation.

You work well independently but value collaboration. You enjoy discussing ideas, sharing work early, and improving your thinking through interaction with researchers from different backgrounds.

You care deeply about rigor and transparency. You surface uncertainty clearly, challenge weak assumptions, and avoid overstating conclusions.

Above all, you are motivated by the opportunity to contribute to ambitious scientific problems that have the potential to change how we understand biological intelligence.

Background We Typically See

We care more about scientific judgment and technical capability than credentials. We typically see:

• Ph.D. in Computer Science, Applied Mathematics, Computational Neuroscience, Machine Learning, or a related field
• Strong track record developing machine learning methods for multidimensional signal, image, or sequence data
• Experience building models across predictive, generative, classification, or regression tasks
• Demonstrated expertise in statistical modeling, predictive analysis, and causal inference
• Experience developing impactful computational tools, models, or software systems
• Proficiency with Python and modern machine learning frameworks such as PyTorch or JAX
• Strong scientific communication and collaborative software development practices

Exceptional candidates may come from different paths. What matters most is the ability to apply machine learning thoughtfully to difficult scientific problems and generate meaningful scientific insight.

What We Offer

We offer competitive compensation and comprehensive benefits, including:

• Unlimited Paid Time Off (PTO), plus 12 company holidays
• Comprehensive medical, dental, and vision coverage
• 100% company-paid short-term disability, long-term disability, basic life insurance, and AD&D insurance
• 16 weeks of fully paid parental leave
• Relocation assistance for qualifying relocations