Video Research Intern: ML-Driven Streaming (Hybrid)

MediaKind; Rennes; Hybrid; EUR 10,000 - 12,000; Full time

Benefits offered by this job

Paid internship

Public Transport: 50% reimbursement

Job summary

MediaKind is offering a Video Research Intern position in Rennes, France, hybrid mode, for six months starting January/February 2027. You will support the Applied Video Research team, work under senior guidance, and own a well-defined project from literature review to experimental results.

You’ll gain hands-on experience applying state-of-the-art video compression and ML techniques, with exposure to both open-source and proprietary tools, plus mentorship and potential for publication or patent

Qualifications

• Enrolled in a Master’s program or final year of an engineering degree in a related field.
• Sign internship agreement (convention de stage) with school/university is mandatory.
• Hands-on experience in ML/DL on image or video problems with Python and a DL framework.

Responsibilities

• Conduct literature reviews and summarize findings for the team.
• Run experiments with codecs, models and datasets.
• Build and maintain evaluation, benchmarking and experimentation tools.
• Develop proofs of concept under guidance from senior researchers.
• Document results in technical reports and presentations.

Skills

Machine learning

Python

PyTorch

TensorFlow

Research

Critical reading

Education

Master’s program or final year of engineering

Tools

x264/x265

SVT-AV1

VVenC

VTM

FFmpeg

Linux

Git

C/C++

Job description

MediaKind is offering a Video Research Intern position in Rennes, France, hybrid mode, for six months starting January/February 2027. You will support the Applied Video Research team, work under senior guidance, and own a well-defined project from literature review to experimental results.

You’ll gain hands-on experience applying state-of-the-art video compression and ML techniques, with exposure to both open-source and proprietary tools, plus mentorship and potential for publication or patent