MorphLabs is hiring its founding research scientist in New York to own the decoders that map EEG to intended movement, working directly with the CTO on transfer learning and closed-loop control.
MorphLabs is building a non-invasive brain interface that turns movement intent into action, starting with continuous, intuitive prosthetic control for upper-limb amputees by decoding movement intent from EEG in real time, without surgery or muscle sensors. It is a small pre-seed team in New York that builds its own headset, electrodes and amplifier, develops hardware and models together, and has an IRB in place covering healthy participants and patients.
About the role
The central problem at MorphLabs is learning a mapping from EEG to intended movement that holds up across people and over time. Those models drive prosthetic and robotic hands and on-screen cursors. The founding research scientist owns that research from first question to result, reporting to the CTO.
What you will do
- Transfer learning across subjects, sessions and devices: build decoders that adapt to new users from minimal calibration data and stay robust to drift.
- Self-supervised pretraining: develop objectives and representations on a growing, largely unlabeled EEG corpus and measure downstream transfer to real control.
- Online closed-loop decoding: move from offline benchmarks to causal, low-latency decoders that adapt during use, and study why offline accuracy does not predict online performance.
- Paradigm design: design the tasks and data-collection protocols that produce informative training data.
- Evaluation and benchmarking: define the metrics, held-out splits and benchmarks used to judge decoders.
- Inform the hardware: set signal-quality standards with the hardware team and help validate the custom headset against third-party systems.
Who they are looking for
A senior individual contributor. A PhD or a background in neuroscience or BCI is not required.
- Research you can point to, with a clear account of what you did yourself.
- Models you built and trained on noisy time series, with the ability to explain architecture and objective choices.
- Models that kept working when the data shifted across people, sessions, devices or environments.
- Rigorous, reproducible work: fair baselines, ablations, and test sets that hold out whole sessions or people.
- Strong candidates may also have biosignal or non-invasive BCI experience, real-time or closed-loop systems experience, or pretraining results that carried to a downstream task.
Equity of 0.2% to 2%, unlimited PTO, 401(k) with matching, health insurance, visa sponsorship and a relocation stipend. Target start December 2026.