Post-Doctoral Researcher
Core
Develop techniques to build a foundation model of human neurotypes by integrating cognitive task performance, conversational text, and biometric data using self-supervised representation learning.
Role type
Postdoctoral Researcher (Multimodal Machine Learning)
Builds
A continuous latent manifold of individual differences representing human neurotypes
Domain
Neuroscience + Multimodal Machine Learning
Deliverable
production ML models
Required skills
Self-supervised learning, Multi-modal modeling, Representation learning, Statistical evaluation, Python, Deep learning frameworks, High-performance computing
Preferred skills
Natural language processing, Cognitive modeling, Privacy-aware methods
Technologies
Python, GPU clusters
Responsibilities
Develop tokenization techniques and self-supervised learning methods to integrate diverse data streams; Train models utilizing novel multi-modal architectures to project data into a unified latent space; Design rigorous evaluations focusing on participant-level data separation, model robustness, missing modalities, and generalization; Scrutinize models for bias, unsupported sensitive inferences, and stability; Collaborate with software engineers on data preprocessing pipelines and model integration within cloud infrastructure.
