Über
Type Unpaid (volunteer research Volunteership)
Time Commitment 15 hours per week
Location Remote
What You’ll Work On Develop and refine model architectures for three modalities: cognitive test embeddings, voice test embeddings, and vascular biomarkers.
Implement modules including:
Learned encoders and latent-space disentanglement
R-GAT graph encoders for cognitive subdomains
HuBERT-based audio embedding pipelines
Custom latent fusion VAEs using product-of-experts rules
Multimodal diffusion-transformer and LSTM hybrid networks for time-series trajectory prediction
Bayesian heads for uncertainty estimation
Collaborate with the engineering and research teams to containerize experiments and prepare inference-ready MVP modules.
Contribute to documentation, reproducibility, and validation of models on synthetic and real datasets.
Requirements
Proficiency in Python and PyTorch.
Must be incoming or currently enrolled in a Bachelors/Associates (or any equivalent) Program at a licensed post-secondary institution.
Experience implementing sequence models (LSTM/GRU), Transformers, and VAEs.
Familiarity with graph neural networks (GAT/R-GAT) and multimodal fusion methods.
Comfortable reading and implementing from academic research papers.
Strong debugging, version control (Git/GitHub), and documentation skills.
Availability of 15 hours per week with consistent progress updates.
Preferred
Prior experience working on multimodal or neuro-related ML projects.
Understanding of probabilistic modeling, diffusion models, or Bayesian methods.
Exposure to MLOps practices (Docker, experiment tracking, etc.).
Benefits
Collaborate on a real-world neuro-AI project with technical depth and publication potential.
Gain experience in full-stack ML research from architecture design to deployment.
Earn a detailed technical reference and recognition upon successful completion.
How to Apply Interested candidates can apply via Linkedin Quick.
Applications are reviewed on a rolling basis. Please include your GitHub, portfolio, and 1–2 examples of relevant ML work or projects.
#J-18808-Ljbffr
Sprachkenntnisse
- English
Hinweis für Nutzer
Dieses Stellenangebot stammt von einer Partnerplattform von TieTalent. Klick auf „Jetzt Bewerben”, um deine Bewerbung direkt auf deren Website einzureichen.