Post-Doctoral Researcher in AI Benchmarking

24 months contract | Belval 

Are you passionate about research? So are we! Come and join us

The Luxembourg Institute of Science and Technology (LIST) is a Research and Technology Organization (RTO) active in the fields of materials, environment and IT. By transforming scientific knowledge into technologies, smart data and tools, LIST empowers citizens in their choices, public authorities in their decisions and businesses in their strategies. 

Do you want to know more about LIST? Check our website: https://www.list.lu/

Discover our IT for Innovative Services department: https://www.list.lu/en/informatics/ 

How will you contribute?

Within the Trustworthy AI group, we are looking for a post-doctoral researcher with strong AI benchmarking expertise for a European Defence Funds project. If you want to do research with a strong impact, and have an interest for working alongside world-class defence actors, let us know of your profile.

The ideal candidate would cover a mix of the following points 

[MUST] Definition of AI Metrics and Experimentation

• Semantic segmentation metrics: class- / task-specific and/or instance-level metrics, pixel-level, panoptic, metrics for resilience, and use-cases that target real-time and long-term performance evaluation, model complexity, … 

[NICE] Definition of AI Metrics and Experimentation

• Hardware-specific AI metrics: inference speed, memory usage, power consumption, single/double-precision, performance on GPUs, TPUs, edge, … 

[MUST] Semantic Segmentation Explainability Benchmarking

[MUST] Reproducibility and metrics 

[MUST] Impact of Data on Performance

• GAN-driven augmentation / adversarial synthetic data generation for challenging a model’s performance, domain adaptation (e.g., image-to-image translation, style transfer augmentation) 

 [NICE] Impact of Data on Performance

• Impact of data split and augmentation on model performance, overfitting, sensitivity to class imbalance, transfer-learning and fine-tuning performance
• Data augmentation for testing model resilience & re-assessment: occlusion, jittering, regularization techniques, geometric and/or photometric transformations, inpainting, …
• Temporal and spatial split: concept drift, train vs inference, class-aware split (low-representativeness class benchmarking, long-tail distribution), out-of-distribution generalization  

[NICE] Hyper-model / model sampling, latent model space metric learning

[NICE] Tooling; standardized benchmark protocol(s) and reporting

• Experience with tooling for AI model versioning, experiment tracking, MLOps & orchestration, containerization, model serving, REST APIs
• Reporting and support for post-benchmarking drill-down on results

Our team is active across different technology-transfer areas (AI benchmarking, explainability, embedded/edge AI, …). 

Is Your profile described below? Are you our future colleague? Apply now! 

The scope of the work is on AI benchmarking, while observing applications for semantic segmentation and anomaly detection.

The successful candidate will be responsible with conducting state-of-the-art research in the area of AI benchmarking for semantic segmentation / AI models, publishing in high-end journals and venues, operational work for the implementing metrics, benchmarking pipelines, etc., all in a solution-focused environment. 

The successful candidate is expected to:

• Have a strong ML/AI background, hold a PhD in Computer Science, Mathematics or similar, and be able to showcase a track record in AI benchmarking research.
• Hands-on experience and a track record of projects carried using frameworks like Pytorch and/or Tensorflow is an absolute must.
• Any experience in working with e.g. Kubeflow, Kubernetes / associated tooling, and benchmarking frameworks like MLPerf / others, is a plus while not a strong requirement. 

Operational knowledge and fluency in English is mandatory; French can be a plus while not required nor mandatory.

Any previous experience in a defense-oriented context is a plus while not mandatory; similarly, having been granted an EU Personal Security Clearance in the past can be a strong advantage. 

Your LIST benefits

• An organization with a passion for impact and strong RDI partnerships in Luxembourg and Europe that works on responsible and independent research projects
• Sustainable by design, empowering our belief that we play an essential role in paving the way to a green society
• Innovative infrastructures and exceptional labs occupying more than 5,000 square metres, including innovations in all that we do
• An environment encouraging curiosity, innovation and entrepreneurship in all areas
• Personalized learning programme to foster our staff’s soft and technical skills
• Multicultural and international work environment with more than 50 nationalities represented in our workforce
• Diverse and inclusive work environment empowering our people to fulfil their personal and professional ambitions
• Gender-friendly environment with multiple actions to attract, develop and retain women in science
• 32 days’ paid annual leave, 11 public holidays, 13-month salary, statutory health insurance
• Flexible working hours, home working policy and access to lunch vouchers

Apply online

Please apply ONLINE formally through the HR system. Applications by email will not be considered.

Your application must include:

• A motivation letter oriented towards the position and detailing your experience
• A scientific CV with contact details
• List of publications (and patents, if applicable)
• Contact details of 2 references 

Application procedure and conditions

We kindly request applicants to provide their nationality for statistical purposes only, as part of our commitment to promoting diversity and ensuring equal opportunities in our workforce. This information will be kept confidential and will not be used for any discriminatory purposes.

LIST is dedicated to maintaining an inclusive work environment and is an equal opportunity employer. We are committed to attracting, hiring, and retaining a diverse workforce. All applicants will be considered for employment without discrimination based on national origin, race, colour, gender, sexual orientation, gender identity, marital status, religion, age, or disability.

Applications will be continuously reviewed until the position is filled. An assessment committee will thoroughly evaluate applications, adhering to guidelines designed to ensure equal opportunities.

The primary criteria for selection will be the alignment of the applicant's existing skills and expertise with the requirements mentioned above.