Research Scientist, Data Analysis

The Role

This is the person who turns our experiments into results.

Arbor is generating novel datasets. In humans, we image deep brain structures with functional ultrasound, and we will soon be adding ultrasound neuromodulation. Those studies are longitudinal and causal: intervene, image the circuit response, track the symptom, and repeat in the same person over months. In animals, we combine this bidirectional ultrasound with electrophysiological stimulation and recording, giving us causal measurements across a wide range of spatial and temporal scales, in the same subjects over time.

Your primary responsibility is the analysis of those data and the tooling around them, working with a highly quantitative experimental team from the design stage onward. Because these studies will be longitudinal and closed-loop, there will also be a need for optimal experiment design, online analysis implementations, and closed-loop control.

We're hiring at multiple levels and will calibrate scope and title to your experience: at the earlier end, you'd own the analysis work in close collaboration with the scientific team; with more experience, you'd set the computational direction for the organization.

We're also still a small, startup-like organization, so everyone fills multiple roles, including hands-on science and the work that gets science done.

What You’ll Do

Own analysis of Arbor's experimental data, including functional ultrasound imaging and electrophysiology, along with the tooling that supports it

Work with the experimental team from study design through interpretation, so we answer the highest-priority questions

Develop image processing tools: registration and alignment, artifact handling, and whatever else the data demands

Help design algorithms for intervention optimization and adaptive experiment design

Develop online and real-time analysis methods to support closed-loop experiments

Build methods for integrating data across modalities and scales

Contribute to high-level scientific decisions about where the organization is going

Not all of this is one person's job. Where you start depends on your strengths and on what the program needs most.

Who We’re Looking For

The successful candidate will have:

• A strong quantitative background: a PhD, or comparable training in an equally demanding environment such as industry

• Real depth in several of: signal processing, statistics and inference, machine learning, medical or brain imaging analysis, control theory, computational neuroscience; we don't expect all of them

• Strong coding skills

• A significant body of independent work you can point to

• The judgment to tell a problem that can be solved with standard statistical methods from one that requires a more sophisticated approach – and the discipline to prefer the simpler answer

• A self-starter with the skills and instinct to find the shortest effective path to the scientific goal, rather than the most interesting one

• Directness without friction – you say what you think, ask when you don't know, and aren't looking to compete with your colleagues

Nice to have: experience in a professional software development environment; a public record of algorithm or code development on GitHub or similar; experience with ultrasound, neural, or medical imaging data; prior work in a research setting where you had to build the tooling yourself.

You don't need a neuroscience background. You do need to want one. Some of the best people to have worked on these problems originally came from another field.