Applied Sciences Intern

About the role

As an Applied Sciences Intern at Microsoft you will work on cutting‑edge research projects that directly influence product quality and business outcomes. You will collaborate with researchers, engineers and product teams to translate real‑world challenges into machine‑learning research problems and prototype intelligent solutions.

Key responsibilities

• Design, implement, and evaluate machine learning and AI solutions for real‑world product scenarios.
• Analyse and improve the performance of advanced algorithms on large‑scale datasets.
• Translate business and product challenges into experimental frameworks and rigorous experiments.
• Develop, train, fine‑tune, and evaluate models including Large Language Models (LLMs) and Small Language Models (SLMs).
• Build scalable prototypes and AI‑powered systems ready for production deployment.
• Prepare, clean, and analyse large datasets while ensuring data quality and integrity.
• Develop innovative approaches to model evaluation, benchmarking, and performance optimisation.
• Document research findings and contribute to technical reports, publications, patents or open‑source initiatives.

Required profile

• Currently pursuing a PhD in Computer Science, Artificial Intelligence, Machine Learning, Statistics, Mathematics, Electrical Engineering, Computer Engineering, Econometrics or a related technical field.
• At least one semester/quarter remaining after the internship.
• Research experience in Machine Learning, Deep Learning, Natural Language Processing, Computer Vision, Generative AI or related areas.
• Demonstrated ability to formulate hypotheses, conduct independent research and communicate technical findings effectively.

Required skills

• Strong programming skills in Python.
• Experience with machine learning frameworks such as PyTorch or TensorFlow.
• Experience designing experiments, analysing results and applying statistical methods.
• Experience working with large datasets and building end‑to‑end machine learning pipelines.