We expect the intern to display independence and maturity as a researcher.
The intern is expected to design, implement, and evaluate our methodology, according to the best practices of the field and inspired by, e.g., their previous research experience, new insights they glean from related literature, collaborative team discussions, and in-depth discussions with the supervisor(s).
To be successful, the intern must understand and have experience in dealing with relevant challenges in representation learning and robotics research, such as:
leveraging foundation models for robot decision-making and control,
determining regularization strategies and self-supervisory tasks for efficiently learning effective multimodal representations,
pursuing model robustness and generalizability to distribution shifts (e.g., unseen task-execution settings, simulation-to-real gap, policy transfer across different robot morphologies),
dealing with the practicalities related to implementing neural policies (e.g., optimisation tricks, multi-machine/multi-GPU training),
conducting model performance characterization + error analysis (determining informative ablations and baselines, visualizing and interpreting trajectories, visualizing and inspecting learned representations, identifying dataset biases), etc.
Requirements & Skills:
Strong background in machine learning, with particular emphasis on multimodality and/or representation learning
Strong background in Robotics or Embodied Artificial Intelligence
Extensive experience in from-scratch neural model implementation, e.g., using PyTorch
Extensive experience with data analytics toolkits, such as numpy, pandas, and scikit-learn
Extensive experience in implementing, training/fine-tuning, and evaluating the performance of CV models, NLP models, policies, etc.
Extensive experience in software development in Python on Linux-based systems
Extensive experience in training neural models on multi-machine or multi-GPU setups
Extensive publication history in top conference venues; is a mature researcher
(Preferred) Experience in leveraging Large Language Models, Vision-Language Models, and/or foundation models that are grounded with other modalities (e.g., audio, haptics, etc.)
(Preferred) Strong theoretical background in AI/ML/Robotics topics, e.g., representation learning, transfer learning, reinforcement learning, learning from demonstrations, safe learning, etc.
Your degree level: pursuing doctoral degree, or current post-doctoral researcher
• Your major: Computer Science, Electrical & Computer Engineering, Statistics, or related
