Call for Internships

Minhyuk Sung, KAIST, Summer 2022

Deadline: May 15 (Sun), 23:59 KST.
Application Form: Google Form Link

We’re looking for undergraduate interns who are interested in building cutting-edge machine learning technologies for 3D geometric data. ML for 3D is a rising topic in multiple research communities including computer vision, computer graphics, and robotics — for example, see the paper statistics of CVPR 2020 and 2021 here. `3D’ is ranked in the top 3 in a row. During the internship, you will work on your own research project with a topic in this field and learn and discuss with us 1) how to find a good research topic, 2) how to develop your research idea, 3) how to implement and prototype the idea, and finally (and possibly), 4) how to demonstrate the research outcomes in a paper.

Research topics include, but are not limited to:

• Neural rendering
• Generative models for 3D data
  • Score-based diffusion models for 3D data
  • Generative models for textures
• Language-to-3D
• Discovering structure in 3D
• Learning generalizable 3D structure
• 3D deformation
• Unsupervised 3D segmentation
• 3D shape representations for completion and denoising
• Learning with 3D graphs
• Learning for human poses
• Geometry of feature spaces

For reference, check out

• CS492(A): Machine Learning for 3D Data course webpage, and
• My webpage.

Requirements

• We’re looking for students who have experience in developing any deep learning techniques (either with PyTorch or TensorFlow).
• Candidates who have taken or audited the CS492(A): Machine Learning for 3D Data course will be preferred (but not required). All potential candidates will be asked to finish following the course materials before starting the internship.
• Interns will be required to be physically present in our lab during the internship.
  Location: KAIST, N1 Bldg, Rm 622, 291 Daehak-ro, Yuseong-gu, Daejeon, Korea 34141.
• Candidates who can extend the internship after the summer will be preferred. (Remote work can be considered for the extension.)

Conditions

• Interns will be expected to commit to the internship for full-time at least 8 weeks this summer. The starting and the end dates are flexible, but the desired start and end dates are:
  Start date: Jun 27 (Mon)
  End date: Aug 26 (Fri)
  (including one week break during the internship)
• Interns will be paid and provided with a desk in our lab and computing resources (access to either KCloud or our own GPU servers).

How to Apply

Deadline: May 15 (Sun), 23:59 KST.

Please submit the following application form:
Google Form Link

You’ll need to prepare the followings:

• your CV/resume (including your previous experience in research, internship, or course projects, and your programming skills. Not required but recommended).
• a copy of your transcript, and
• a short description of why you’re interested in this internship and what you expect to do during the internship.

FAQ

• Q. Can non-KAIST students apply?
  A. Yes.

• Q. Can non-KAIST students stay in a KAIST dormitory during the internship?
  A. Yes.

• Q. Can people who are not enrolled in any Korean institutions apply?
  A. Yes, but we will NOT be able to pay you for the internship in this case.

More References

If you want to learn more about ML techniques for 3D geometric, please check out this website:
3D Machine Learning Archive

also these great lecture series (you can watch previously recorded lectures):
3DGV Seminar
Toronto Geometry Colloquium