Meta AR/VR Job | Research Scientist Intern, XR Data Ingestion Systems (PhD)
Job（岗位）: Research Scientist Intern, XR Data Ingestion Systems (PhD)
Type（岗位类型）: Computer Vision | Engineering, Research
Citys（岗位城市）: Burlingame, CA
Date（发布日期）: Before 2021-12-14
Summary（岗位介绍）
Our team is responsible for maintaining the quality and infrastructure of petabyte scale video and image anonymization.
Pseudonymization is replacing actual human faces in training data with synthetic faces. We need to invest in this area since training data for anonymization models must not not contain any PII. We have an option to either blur the training data and hope the net recognizes body parts to blur the face OR we pseudonymize the data thus eliminating any privacy concerns. The internship will objectively measure both approaches.
Impact:
Currently, the anonymization model teams are severely impacted because of strict privacy requirements for our training data. With a stagnant test+training dataset we are lacking the ability to improve the model. We’d like to record all known misses of the model and use it to improve training data – but can’t today.
If pseudonymization or blurred faces proves to only mildly affect precision and recall (mean average precision across the two), we will be able to unlock the ability to have very accurate anonymization models that work for bespoke domains.
Qualifications（岗位要求）
Currently has, or is in the process of obtaining, a PhD degree in Computer Science, Electrical Engineering, or Electrical and Computer Engineering in the field of computer vision, machine learning, computer graphics, or robotics.
3 years experience with systems building in C++ and Python.
2 years experience with modern deep learning frameworks like PyTorch.
Knowledge in computer vision, machine learning, image processing, object recognition or object tracking.
Proven track record of achieving significant results as demonstrated by publications in top computer vision conferences (e.g., CVPR, ICCV, ECCV, or, SIGGRAPH) or journals (e.g., IEEE Transactions on Pattern Analysis and Machine Intelligence, International Journal of Computer Vision, Pattern Recognition, or IEEE Transactions on Image Processing).
High levels of creativity and quick problem solving capabilities.
Must obtain work authorization in country of employment at the time of hire, and maintain ongoing work authorization during employment.
Description（岗位职责）
Comparative analysis of how blurring faces and pseudonymization in training data affect precision and recall on a golden dataset.
Create a system (or use handcrafted data) that can replace instances in an image with synthetic faces.
Additional Requirements（额外要求）
Demonstrated software engineering experience via an internship, work experience, coding competitions, open source contributions, or research.
Ability to communicate complex research in a clear, precise, and actionable manner.
Intent to return to degree-program after the completion of the internship.
