The Opportunity

We are looking for a master thesis student at our R&D department, would like to join our multinational organization with lots of development opportunities. Our team is dedicated to creating a future where innovative digital technology brings cleaner energy to people around the world. If this opportunity excites you, we encourage you to apply!

How you'll make an impact

• Determine a suitable array of time series machine learning models, such as recurrent neural networks (RNNs), long short-term memory (LSTM) networks, and possibly transformer-based models. Implement these models modularly to enable their deployment on a variety of hardware types.
• Configure and optimize the three processor environments CPU, GPU, and NPU by employing industry standard frameworks like TensorFlow and PyTorch, ensuring that each setup is tailored for effective benchmarking.
• Defining performance metrics and benchmarks requires to pinpoint key indicators such as training time, inference speed, energy consumption, and memory usage. We will benchmark each hardware type across chosen applications, with a keen focus on comparing processing times for small, medium, and large datasets, and evaluating the scalability of each processor type.
• Recommendations and guidelines based on the findings, offer practical guidelines on hardware selection for different types of time series applications. Discuss trade-offs between performance, cost, and scalability.

• Bachelor's degree in computer science, electrical engineering, data science, or a related field.
• Proficiency in python, with experience in deep learning libraries (e.g. TensorFlow, PyTorch) and familiarity with hardware configuration for machine learning tasks.
• Knowledge of time series analysis and machine learning algorithms, particularly recurrent and convolutional neural networks.
• Basic understanding of CPUs, GPUs, and NPUs, as well as parallel computing principles.
• Prior experience in machine learning model deployment, optimization, or hardware benchmarking would be beneficial.