Machine Learning Researcher

About the company Orbit is building the foundational AI Infrastructure for emotions. A read interface to human emotions will be the most consequential data layer, enabling empathetic AI and precision mental healthcare. It will fundamentally transform how we interact with each other and technology. Backed by founders and operators of top AI labs, consumer hardware, pharma, and enterprise companies, and venture-backed. About the team we are building We’re building a generational founding team which is truly full-stack - from neural sensors to complex models. If you want to work on deep technological problems and help pioneer the future of NeuroAI, this is the place for you. Projects have opportunities for a high degree of autonomy and demand intense, fast-paced learning. About you Strong Python programming ability with a track record of building and iterating quickly on ML models Skilled at designing robust evaluation pipelines and benchmarks for novel architectures Comfortable working with large, noisy, or unconventional datasets Experience with data preprocessing, labeling, and exploratory analysis Familiarity with multi-modal architectures and integrating heterogeneous data sources Excited to learn neuroimaging and neuroscience context (we will support you in getting up to speed) Preferred Qualifications/Experience MS or higher in Computer Science, Electrical Engineering, Applied Mathematics, or related STEM field; exceptional self-taught researchers also considered 3+ years of applied ML research or development experience, or equivalent depth through publications, projects, or startup work Publications in top ML or domain-specific journals/conferences Prior work with multi-modal data (e.g., imaging + time-series, text + audio, etc.) Experience designing and running experiments with novel model architectures Familiarity with PyTorch, TensorFlow, or JAX Hands-on experience in startups, small research groups, or similarly fast-moving environments Nice-to-have Experience with biomedical, neuroimaging, or other high-dimensional sensor data Background in signal processing for time-series or imaging data Knowledge of model training at scale (distributed, mixed precision, large datasets) Experience with semi-supervised or self-supervised approaches