Internship | Task-specific AI Model Selection Based on Energy and Quality Tradeoffs

Internship | Task-specific AI Model Selection Based on Energy and Quality Tradeoffs

What will be your role?

As AI systems become increasingly embedded in diverse domains—from healthcare to finance—the need for transparent evaluation of their sustainability grows. Most AI research to date primarily benchmarks performance (accuracy, F1 score) at a broad task or system level, frequently ignoring the real energy impact of deploying such models for highly specific or granular workflows.

For example, while general speech-to-text benchmarks are widespread, little attention is paid to the energy and performance variance when models are used for domain-specific sub-tasks (e.g., transcribing clinical terminology in operating rooms versus everyday speech). In practice, optimizing for these specific tasks—including fine-grained model selection and adaptive deployment—can lead to significant reductions in resource usage, and thus, emissions.

At TNO, sustainability of AI systems is part of our mission to make AI systems more responsible. We seek to investigate new methods that systemically measure and compare both performance and energy consumption of models when executed on narrowly-defined tasks. By correlating accuracy with energy usage at the task-level, we aim to develop actionable guidelines and tools for practitioners to make environmentally conscious decisions in model selection and deployment.

Examples of fitting research questions:

• How can we experimentally assess and visualize the energy cost of different AI architectures when applied to a well-defined micro-task?
• Can we develop methodologies to benchmark energy-performance tradeoffs for models dynamically, as task complexity or input characteristics change?
• How do model and inference-serving design choices (e.g. batched inference, speculative decoding, kv caching, quantization of weights, flashattention, ...) influence energy consumption and accuracy tradeoffs?

What will you be doing?

As a graduation project candidate, you will design and implement experimental setups for energy measurement of AI models at the task level. You will explore literature on existing model benchmarking, energy profiling, and sustainability metrics. By developing tools and protocols, you will run controlled experiments on various tasks (such as entity extraction, domain-specific translation, or anomaly detection) and document both performance and energy data. You will have the opportunity to analyze and compare results across architectures, model sizes (LLMs, SLMs, classifiers) and deployment scenarios. Ultimately, your work will contribute to TNO’s suite of sustainable software and AI best practices. You will be supervised by experts in the Green IT and Data Science groups.

What we expect from you

We are looking for a student, who is interested in the field of (generative) AI and software engineering, who is interested in research and enjoys tackling technically challenging problems. The student should be able to work unsupervised and have a structured approach to accomplish their goals in their research proposal. Furthermore, the student should be comfortable working in a team and be willing to learn more about the domain and the field.

You are following a master’s study in a relevant field, such as artificial intelligence or computer science, etc. You possess good programming skills. You are currently studying within the Netherlands or other EU country and be willing to relocate to the Netherlands. You have good communication skills in English.

What you'll get in return

You want an internship opportunity on the precursor of your career; an internship gives you an opportunity to take a good look at your prospective future employer. TNO goes a step further. It’s not just looking that interests us; you and your knowledge are essential to our innovation. That’s why we attach a great deal of value to your personal and professional development. You will, of course, be properly supervised during your work placement and be given the scope for you to get the best out of yourself. Furthermore, we provide:

• A highly professional, innovative internship environment, within a team of top experts.
• A suitable internship allowance (615 euro for wo-, hbo- and mbo-students, for a full-time internship).
• Possibility of eight hours of free leave per internship month (for a full-time internship).
• Use of a laptop.
• An allowance for travel expenses in case you don’t receive an OV-card.

TNO as an employer

At TNO, we innovate for a healthier, safer and more sustainable life. And for a strong economy. Since 1932, we have been making knowledge and technology available for the common good. We find each other in wonder and ingenuity. We are driven to push boundaries. There is all the space and support for your talent and ambition. You work with people who will challenge you: who inspire you and want to learn from you. Our state-of-the-art facilities are there to realize your vision. What you do at TNO matters: impact makes the difference. Because with every innovation you contribute to tomorrow’s world.

Contact

E-mailadres: zef.wolffs@tno.nl