Colour Imaging (Research Engineer)- Finland

microTECH Global LTD; Finland; Remote; EUR 70,000 - 100,000; Full time

Job summary

Huawei Finland is seeking a highly capable AI researcher to develop models for computational photography, image/video enhancement, color processing, and ISP pipeline algorithms for mobile platforms.

You will prototype in PyTorch or TensorFlow, build synthetic data pipelines, identify failure cases, and push NPU-compatible optimizations to enable rapid integration into devices. A PhD or MSc nearing PhD is required, with strong Python and DL libraries.

Qualifications

• PhD or MSc studying towards PhD in computer vision, machine learning, signal/image processing, or associated discipline
• Deep understanding of the theory behind the machine learning, deep learning methods and imaging technologies.
• Proven ability to build, train, fine-tune and assess deep learning models for vision tasks such generative AI, image/video restoration, 3D-vision and multimodal learning etc.
• Strong competency in Python and deep learning libraries like Pytorch and Tensorflow, experience in ONNX and NPU-compatible model optimization is a plus
• Familiar with camera ISP, mobile imaging color processing algorithms and traditional image processing algorithms
• Ability to navigate and innovate in rapidly evolving technical landscapes with ambitious and often ambiguous goals.
• Basic understanding of perceptual image quality, human visual system color processing, and chromatic adaptation mechanisms
• Big advantage for contestants from international competitions in computer science, mathematics, or physics (e.g. IOI, IMO, ICPC, IPhO, IChO)

Responsibilities

• Research and develop AI models for computational photography, image/video enhancement, color processing, ISP pipeline algorithms, spanning architectures such as vision-language models, diffusion image/video transformer models, etc.
• Identify data requirements, guide collection and annotation efforts, and build high-quality synthetic data creation pipeline that enable reliable model training
• Identify model failure cases and develop targeted solutions to improve performance in critical real-world scenarios
• First implement research ideas, prototyping in Pytorch or Tensorflow, and then continue with NPU compatible model optimization for rapid integration into Huawei devices
• Partner with other algorithm and hardware imaging teams to define key problems and deliver effective solutions
• Stay on top of the latest AI algorithms and imaging technologies by proactively exploring the latest and greatest in the field, and refine those towards practical solutions for mobile imaging use cases
• Research, design and implement new color and spectral imaging algorithms for reaching the perceptual targets

Skills

Python

PyTorch

TensorFlow

Deep learning

Vision models

Education

PhD or MSc toward PhD in CV/ML

Tools

ONNX

NPU optimization

Job description

This will include learning from and working with world class leading competence on mobile imaging sensors, spectral sensing and emerging novel photon capturing solutions. You will develop related image and video processing algorithms, help define problems, agree on clear performance metrics, and deliver solutions that perform reliably on real-world data.

Job Descriptions

• Research and develop AI models for computational photography, image/video enhancement, color processing, ISP pipeline algorithms, spanning architectures such as vision-language models, diffusion image/video transformer models, etc.
• Identify data requirements, guide collection and annotation efforts, and build high-quality synthetic data creation pipeline that enable reliable model training
• Identify model failure cases and develop targeted solutions to improve performance in critical real-world scenarios
• First implement research ideas, prototyping in Pytorch or Tensorflow, and then continue with NPU compatible model optimization for rapid integration into Huawei devices
• Partner with other algorithm and hardware imaging teams to define key problems and deliver effective solutions
• Stay on top of the latest AI algorithms and imaging technologies by proactively exploring the latest and greatest in the field, and refine those towards practical solutions for mobile imaging use cases
• Research, design and implement new color and spectral imaging algorithms for reaching the perceptual targets

Qualifications and detailed requirements

• PhD or MSc studying towards PhD in computer vision, machine learning, signal/image processing, or associated discipline
• Deep understanding of the theory behind the machine learning, deep learning methods and imaging technologies.
• Proven ability to build, train, fine-tune and assess deep learning models for vision tasks such generative AI, image/video restoration, 3D-vision and multimodal learning etc.
• Strong competency in Python and deep learning libraries like Pytorch and Tensorflow, experience in ONNX and NPU-compatible model optimization is a plus
• Familiar with camera ISP, mobile imaging color processing algorithms and traditional image processing algorithms
• Ability to navigate and innovate in rapidly evolving technical landscapes with ambitious and often ambiguous goals.
• Basic understanding of perceptual image quality, human visual system color processing, and chromatic adaptation mechanisms
• Big advantage for contestants from international competitions in computer science, mathematics, or physics (e.g. IOI, IMO, ICPC, IPhO, IChO)