About InterDigital
InterDigital is a global research and development company specializing in wireless communications, video, artificial intelligence (AI), and related technologies. We develop foundational innovations that enable connected, immersive experiences across communications and entertainment products and services.
Our technologies are licensed worldwide to manufacturers of wireless devices, consumer electronics, Internet of Things (IoT) products, and motor vehicles, as well as to providers of cloud-based services such as video streaming. Since our founding in 1972, our engineers have contributed to advances ranging from the earliest digital cellular systems to 5G and modern Wi-Fi. InterDigital is also a leader in video processing, encoding, and decoding, supported by extensive AI research across wireless and video. InterDigital is listed on Nasdaq.
InterDigital is a registered trademark of InterDigital, Inc.
For more information, visit: www.interdigital.com .
Internship Overview
The Video Production Services team develops technologies that improve how video is produced and delivered for modern displays, with particular expertise in high dynamic range (HDR) and wide color gamut. This internship will focus on optimizing an existing neural network that converts standard dynamic range (SDR) video to HDR, then deploying the optimized model on dedicated neural processing hardware.
Technical Approach
The intern will build an end-to-end optimization workflow using techniques such as knowledge distillation and quantization. The optimized model will be benchmarked against the original, deployed to a low-power and compute-constrained target, and assessed for visual quality, output consistency, model size, latency, and runtime efficiency.
Key Responsibilities
Analyze the existing SDR-to-HDR network and establish quality, performance, and hardware constraints.
Select suitable optimization techniques and supporting tools, including knowledge distillation and quantization frameworks.
Develop a reproducible pipeline for model optimization, evaluation, and deployment.
Benchmark the optimized model against the original for visual quality, output consistency, model size, latency, and computational efficiency.
Deploy and validate the optimized model on a dedicated neural processing architecture.
Document the methodology, results, limitations, and recommendations.
Candidate Profile
Required Qualifications
Currently pursuing a master’s degree in computer science, electrical engineering, applied mathematics, or a related discipline.
Proficient in Python and C/C++.
Experienced with a deep-learning framework such as PyTorch or TensorFlow, including neural network training and evaluation.
Able to communicate clearly in written and spoken English.
Autonomous, curious, and rigorous in research and experimentation.
Preferred Experience
Experience with GPU programming or inference optimization tools such as CUDA, TensorRT, ONNX, LiteRT/TensorFlow Lite, or Core ML.
French-language skills.
Technical Focus Areas
Deep learning and neural network optimization
Image and color processing for SDR and HDR video
Knowledge distillation and neural network quantization
Deployment on dedicated neural processing architectures
Model quality and performance evaluation
Deep learning and neural network optimization
Image and color processing for SDR and HDR video
Knowledge distillation and neural network quantization
Deployment on dedicated neural processing architectures
Model quality and performance evaluation
Expected Deliverables
A structured, documented codebase for model distillation, quantization, evaluation, and deployment.
An optimized SDR-to-HDR model deployed on a dedicated neural processing architecture.
A comparative assessment of the original and optimized models covering visual quality, output consistency, model size, latency, and efficiency.
Actionable recommendations for future optimization and deployment work.
Location: Rennes, France
Mentors: To…