Deep Learning Intern, Model Optimization

Role

In this role, you will focus on optimizing state-of-the-art computer vision and deep learning models for deployment in industrial settings. You will research and apply techniques to compress complex models, making them smaller and faster while maintaining high accuracy. You will work on implementing knowledge distillation and quantization strategies to adapt these models for specific hardware constraints. Throughout your internship, you will design experiments to evaluate performance trade-offs between speed, size, and precision, and you will collaborate with the team to integrate these optimized models into our production software stack.

How your work moves the mission forward

• Accelerate the deployment of advanced perception capabilities by significantly reducing model latency and memory footprint.
• Enable real-time robotic control and decision-making by ensuring complex models run efficiently on constrained edge hardware.
• Identify and validate novel optimization methods that define the future standard for our machine learning pipelines.

Skills you will need to be successful

• Currently pursuing a PhD or Master’s degree in Computer Science, Electrical Engineering, Robotics, or a related technical field.
• Experience with Python and major deep learning frameworks (e.g., PyTorch, JAX, or TensorFlow).
• Theoretical understanding and practical experience with model compression techniques, specifically knowledge distillation and quantization.
• Experience training and evaluating deep learning models for computer vision tasks (e.g., Transformers, CNNs).
• Ability to analyze experimental data to diagnose performance bottlenecks and report findings clearly.