Overview
Research Internships at Microsoft provide a dynamic environment for research careers with a network of world-class research labs led by globally-recognized scientists and engineers, who pursue innovation in a range of scientific and technical disciplines to help solve complex challenges in diverse fields, including computing, healthcare, economics, and the environment.
We are seeking Research Interns to advance how foundation models and artificial intelligence (AI) systems acquire, represent, and reuse knowledge. Today's models hold knowledge implicitly in their parameters and reason explicitly in text. The agent systems built on them gain large amounts of execution experience that they rarely learn from. This Research Internship explores both sides of the problem:
Knowledge in the model: implicit, parameterized knowledge representations and latent reasoning that make large foundation models more trustworthy, compositional, and efficient. This includes how neural representations can complement or replace explicit, symbolic knowledge.
Knowledge from experience: recursive self-improvement (RSI) for AI agent systems. Agents learn from execution experience and progressively improve both their task-solving workflows and the strategies they use to generate future improvements. The focus is on long-horizon enterprise workflows, and the goals are better task quality, reliability, and token efficiency while keeping system changes verifiable and controlled.
Research Interns will contribute to one or both thrusts, depending on their background and research interests. Expected outcomes include research publications and reusable open-source components.
Responsibilities
Research Interns put inquiry and theory into practice. Alongside fellow doctoral candidates and some of the world’s best researchers, Research Interns learn, collaborate, and network for life. Research Interns not only advance their own careers, but they also contribute to exciting research and development strides. During the 12-week internship, Research Interns are paired with mentors and expected to collaborate with other Research Interns and researchers, present findings, and contribute to the vibrant life of the community. Research internships are available in all areas of research, and are offered year-round, though they typically begin in the summer.
Research Interns will work in one of the following research areas, depending on their background and interests.
Knowledge Representations and Implicit Reasoning
Study trustworthy, implicit knowledge representations and their relationship to symbolic knowledge and explicit Chain-of-Thought (CoT), including latent reasoning and looped Transformers.
Develop self-supervised methods for reusable, compositional knowledge representations across structured reasoning, semantic transformations, and decision-making.
Design and train foundation models for efficient in-context learning, adaptation, and generalization to unseen tasks, drawing on tabular foundation models and implicit neural reasoning.
Explore scalable synthetic data, self-supervised objectives, and large-scale pretraining for continuous knowledge acquisition and efficient reuse.
Build prototypes and rigorously evaluate representation quality, compositional generalization, sample efficiency, and inference efficiency.
Recursive Self-Improvement for Agent Systems
Analyze execution traces, tool interactions, and evaluation feedback to diagnose recurring failures, redundant reasoning, unnecessary tool calls, and ineffective context use.
Develop targeted, verifiable improvements to agent harnesses, memory, and tool-use strategies, using prior optimization experience to make subsequent improvements more efficient and controlled.
Evaluate long-horizon enterprise, computer-use-agent, and multimodal reasoning workflows for task success, token consumption, latency, and consistency across runs; assess transfer to unseen tasks and progress across self-improvement rounds.
Across both areas, publish findings in top-tier AI conferences and journals, release reusable open-source components, and collaborate with researchers and engineers to translate advances into practical AI systems.
Required Qualifications
Currently enrolled in a Doctorate Degree program in Computer Science, Artificial Intelligence, Machine Learning, or a related STEM field.
Other Requirements
Research Interns are expected to be physically located in their manager’s Microsoft worksite location for the duration of their internship.
In addition to the qualifications below, you’ll need to submit a minimum of two reference letters for this position as well as a cover letter and any relevant work or research samples. After you submit your application, a request for letters may be sent to your list of references on your behalf. Note that reference letters cannot be requested until after you have submitted your application, and furthermore, that they might not be automatically requested for all candidates. You may wish to alert your letter writers in advance, so they will be ready to submit your letter.
Preferred Qualifications
Research experience in one or more of the following: self-supervised representation learning, large language models (LLM) pre-/post-training, in-context learning, meta-learning, neural/latent reasoning, AI agents, long-horizon reasoning, or AI systems.
Proficient programming skills in Python and hands-on experience developing and evaluating machine learning research systems, such as LLM-based models or agents.
Familiarity with modern self-supervised learning techniques, including contrastive learning, negative-free learning (e.g., BYOL, VICReg), and representation learning objectives.
Experience with large-scale model pretraining, synthetic data generation, and efficient Transformer architectures.
Interest in knowledge representation, compositional generalization, program synthesis, and the relationship between symbolic and neural approaches to intelligence.
Experience building and evaluating LLM-based agent systems, and familiarity with agent harness design, memory and context management, tool use, or workflow orchestration.
Experience with learning from execution feedback, automated optimization, or reinforcement learning.
Experience publishing research papers as a lead author or essential contributor at top-tier conferences such as NeurIPS, ICML, ICLR, ACL, or related venues.
Ability to conduct independent research, design rigorous experiments, and develop reproducible research prototypes.
Keen interest in advancing general-purpose and specialized AI systems through novel learning paradigms.
Intern - MSR- The typical base pay range for this role across Canada is CAD $9,150 - CAD $10,170 per month.
Find additional pay information here: https://careers.microsoft.com/v2/global/en/canada-pay-information.html
Microsoft accepts applications and processes offers for these roles on an ongoing basis throughout the academic calendar (September - April).
Ce poste sera ouvert pendant au moins cinq jours et les candidatures seront acceptées de façon continue jusqu’à ce que le poste soit pourvu.
This position will be open for a minimum of 5 days, with applications accepted on an ongoing basis until the position is filled.
Microsoft est un employeur offrant l’égalité d’accès à l’emploi. Tous les candidats qualifiés seront pris en considération pour l’emploi, sans égard à l’âge, à l’ascendance, à la citoyenneté, à la couleur, aux congés médicaux ou familiaux, à l’identité ou à l’expression de genre, aux renseignements génétiques, à l’état d’immigration, à l’état matrimonial, à l’état de santé, à l’origine nationale, à un éventuel handicap physique ou mental, à l’affiliation politique, au statut de vétéran protégé ou au statut militaire, à la race, à l’ethnie, à la religion, au sexe (y compris la grossesse), à l’orientation sexuelle ou à toute autre caractéristique protégée par les lois, ordonnances et règlements locaux applicables.
Si vous avez besoin d’aide avec des accommodements religieux et/ou d’un accommodement raisonnable en raison d’un handicap pendant le processus de candidature, apprenez-en plus sur la demande d’accommodement.
Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations and ordinances.
If you need assistance with religious accommodations and/or a reasonable accommodation due to a disability during the application process, read more about requesting accommodations.
Insights from previous hires
Top skills: Adaptability, Anticipation, Algorithms, Advertising, Adoption, Account Management, Campaign Management, Campaign Development, C++, Business Reviews
Common previous titles: Software Engineer II, Technical, Tech Support Engineer, Senior Director Product Marketing, Senior Team Manager, Solution Engineer, Software Engineer, Service Engineer, Recruiter, Technical Program Manager
Free tool / Resume tailor
Tailor your resume for this job
Paste your resume and get 3 to 8 of your own lines rewritten in the words this posting uses, with the reason for each change. The job is already filled in, and you don't need an account.
Job description provided by Microsoft.
View the original posting