OverviewThe Teachable AI Experiences (TAIX) team at Microsoft Research Cambridge (UK) is a multi-disciplinary team that aims to drive technical innovation in human-AI interaction through the lens of inclusion. The team is looking to hire an intern with machine learning expertise in multi-modal models, with a particular focus on image-text models, and model adaptation approaches for low-resource domains. The intern will work with several other interns with other disciplinary expertise to contribute to a team project on a platform that brings the voice of marginalized communities into the evaluation and adaptation of large multi-modal models. Candidates should have a passion for equitable AI and should have deep technical knowledge in state-of-the-art image-text models (e.g. GPT-4Vision, CLIP, Flamingo, BLIP, PaLI, Llava), and expertise in at least one of the following areas: AI fairness, model adaption methods (e.g. in-context learning, PEFTs), AI interpretability\/transparency. They should be able to approach technical problems in a multi-disciplinary way and be passionate about building AI technologies that will ensure the inclusion of marginalised communities. The outcomes of the project may lead to a publication in a relevant conference, integration into a Microsoft product\/ application and\/or societal impact. The internship offers a unique opportunity to have real-world impact and drive state-of-the-art research at the intersection of machine learning and disability communities in collaboration with a multi-disciplinary team. Internship will ideally run from April June as part of a cohort of interns.QualificationsRequired\/Minimum Qualifications:Currently pursuing a PhD in machine learning, deep learning, or a related area. All applicants must be currently enrolled in an educational institution.Demonstrable strong technical understanding of state-of-the-art image-text models (e.g. GPT-4Vision, CLIP, Flamingo, BLIP, PaLI, Llava) and expertise in at least one of the following areas: model adaption methods (e.g. parameter-efficient fine-tuning, in-context learning, few-shot learning\/meta-learning), generative approaches like diffusion, AI fairness, AI interpretability\/transparency. This should be evidenced through publications, demos, course projects in these areas.Demonstrable ability to drive high-quality research insights through publications in top-tier machine learning conferences and journals (e.g. NeurIPS,fICML, ICLR, AAAI, ICCV, CVPR, JMLR).Hands-on experience in implementing and empirically evaluating deep learning approaches in PyTorch.Effective communication skills and ability to work in a collaborative environment.Preferred\/Additional Qualifications:The ability to approach technical problems and design solutions with a multi-disciplinary perspective.Passion for ensuring the inclusion of marginalised communities in AI technologies.Previous experience with working in a multi-disciplinary team with diverse skill sets.Authored\/contributed to open-source code projects (e.g. on GitHub).ResponsibilitiesUndertake cutting-edge research in studying the performance of text-image models, including the development of technical approaches to ensure they perform equally well in all scenarios.Write research code to develop and validate new approaches, or develop novel theoretical and practical insights.Collaborate with a diverse and multi-disciplinary team.Clearly communicate research ideas and results in writing, such as research papers, presentations, or research notes for internal and external audiences.Benefits\/PerksListed below may vary depending on the nature of your employment with Microsoft and the country where you work.Industry leading healthcareEducational resourcesDiscounts on products and servicesSavings and investmentsMaternity and paternity leaveGenerous time awayGiving programsOpportunities to network and connect