It's time to reserve tickets for the most challenging adventure of the summer!
Unique project work, for a period of about 3 months.
After completing the internship, you can join a permanent student position or a full-time group or not.
Outstanding students in engineering / computer science? Your place with us!
Examples of some of the specializations offered:
Using simulated data to train our 3D Geometry DNN.
Research data augmentation by rendering different point of views of inference results.
Assist with creating infrastructure for Single Image 3D inference.
Research different 3D Geometry DNN architectures.
Object detection and segmentation benchmark:
Compare results between different architecture.
"Beat" the State of the are performance.
Develop a parallel multi-gpu training: train huge networks in less than a day instead of weeks.
Develop an auto hyper-parameter tuning algorithm.
"Create" hard examples: find out what are the classifiers limits and develop new interesting example using Generative Adversarial Networks (GAN).
The effective receptive field in ConvNets - Develop an object detector which is robust to different input image size.
Unsupervised boosting - build an algorithm for classifiers output anomaly detection and boosting for improving performance.
Dynamic boosting on cloud - develop a training algorithm and a dedicated database for dynamic boosting in AWS.