{"id":4559601,"url":"https://alion.io/job/microsoft-research-intern-self-improving-secure-ai-agents-microsoft-research","title":"Research Intern - Self-Improving & Secure AI Agents - Microsoft Research","company":{"id":18,"name":"Microsoft","domain":"microsoft.com","url":"https://alion.io/company/microsoft","size_band":"5000+","is_staffing_agency":false,"employer_type":"direct","is_intermediary":false,"listed_via":null,"ats_vendor":"Eightfold","truth_index":{"grade":"A","score":85,"open_postings":569,"ghost_share":0.016,"stale_share":0.518,"repost_share":0.065,"time_to_fill_p50_days":30,"computed_at":"2026-11-18T05:48:20Z"}},"role":"AI/ML","role_family":"AI/ML","seniority":"intern","employment_type":"internship","work_mode":"on_site","remote_scope":null,"remote_scope_basis":null,"remote_working_hours":null,"hiring_geo_confidence":"structured","locations":["Redmond, United States"],"countries":["US"],"hiring_countries":[],"hiring_countries_total":0,"salary":{"min":6810,"max":13480,"currency":"USD","period":"month","gross":null,"usd_annual":161760},"salary_estimate":null,"experience_years_min":null,"visa_sponsorship":false,"relocation_package":false,"has_equity":false,"technologies":[{"name":"AI Agents","optional":false},{"name":"Least Privilege","optional":false},{"name":"Machine Learning","optional":false},{"name":"Post-training","optional":false},{"name":"Synthetic Data","optional":false},{"name":"LLM","optional":true},{"name":"Python","optional":true},{"name":"PyTorch","optional":true}],"status":"live","first_seen_at":"2026-11-17T22:51:43Z","employer_posted_date":"2026-11-18","last_verified_at":"2026-11-18T20:26:38Z","board_verified":true,"closed_at":null,"days_open":1,"trust":{"level":"ok","repost_count":null,"flags":[],"days_open":1},"description":"OverviewResearch 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.\nWe are seeking Research Interns to make long-horizon Artificial Intelligence (AI) agents both more capable and more secure. Agents increasingly act autonomously over long horizons and have access to tools, files, networks, and credentials. Two problems limit how far they can be trusted: they fail in ways that aggregate benchmark scores hide, and they can be manipulated by adversarial inputs. This Research Internship offers two research directions, with Research Interns focusing on one based on their expertise and interests:\nFailure-driven self-improvement: improve long-horizon agents by diagnosing failure modes in their trajectories and synthesizing targeted post-training data that closes specific capability gaps.\nAgent security: develop technologies that make existing agentic systems more secure, including sandboxing, agent permissions, and defenses against novel agent attacks.\nResearch Interns will contribute to one thrust, depending on their background. Expected outcomes include research prototypes, evaluation of pipelines, and publications at top-tier venues.\nResponsibilities\nResearch 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.\nAdditional Responsibilities\nProject 1 - Self-Improving AI Agents\nDevelop and evaluate a research agent that uses post-training experiments to iteratively improve a target model through synthetic data.\nTask decomposition and capability diagnosis. Develop methods for the agent to infer the capabilities required by a target task and use evaluation results to identify which capability gaps limit the model’s performance.\nCapability-targeted data synthesis. Build methods that translate hypotheses about capability gaps into training data, with control over skill coverage, example difficulty, and data quality.\nClosed loop optimization. Build a self-improvement loop that measures the effects of each round of synthetic data. Use these results to let the agent revise its capability hypotheses and adapt its data generation strategy across rounds.\nProject 2 - Secure AI Agents\nDevelop and evaluate security mechanisms for existing AI agent systems.\nSandboxing. Isolate agent execution and control access to files, networks, and credentials.\nAgent permissions. Develop least-privilege authorization, task-scoped permissions, and approval mechanisms for sensitive actions.\nAgent attacks and defenses. Study prompt injection, malicious tool outputs, memory poisoning, and cross-agent attacks in controlled environments. Prototype defenses in an existing agent framework and evaluate their security effectiveness, impact on task performance, and usability.\nAcross both projects, Research Interns will collaborate with researchers to communicate findings to technical audiences and pursue publication at top-tier venues\nQualifications\nRequired Qualifications\nCurrently enrolled in a Doctorate Degree program in computer science, machine learning, artificial intelligence, cybersecurity, or a related field.\nOther Requirements\nResearch Interns are expected to be physically located in their manager’s Microsoft worksite location for the duration of their internship.\nIn 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. \nPreferred Qualifications\nDemonstrated Python programming skills and hands-on experience with PyTorch or a similar machine learning (ML) framework.\nPublications at top-tier ML conferences such as NeurIPS, ICML, ICLR, ACL, or related venues.\nHighly self-driven, with the ability to work independently and collaborate effectively in a fast-moving research environment.\nBackground in one or more systems of security, sandboxing, access control, or Large Language Model (LLM)-based agents.\nDemonstrated mathematical foundations and modeling ability.\nDemonstrated hands-on software engineering skills, demonstrated by independently building tools, small systems, or software projects from scratch. Open-source contributions are a plus.\nHigh-level awards or equivalent achievements in competitive programming, mathematics, or algorithmic competitions such as ICPC.\nInterest in security research and ability to design experiments, analyze results, and communicate findings.\nThe base pay range for this internship is USD $6,810 - $13,480 per month. There is a different range applicable to specific work locations, within the San Francisco Bay area and New York City metropolitan area, and the base pay range for this role in those locations is USD $8,920 - $14,610 per month.\nCertain roles may be eligible for benefits and other compensation. Find additional benefits and pay information here: https://careers.microsoft.com/us/en/us-intern-pay\nMicrosoft accepts applications and processes offers for these roles on an ongoing basis throughout the academic calendar (September - April).\nThis position will be open for a minimum of 5 days, with applications accepted on an ongoing basis until the position is filled.\nMicrosoft 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.","description_format":"text","description_chars":7144,"description_truncated":false,"requirements":{"experience_years_min":null,"management_years_min":null,"team_size_min":null,"manages_managers":false,"education":{"level":"phd","optional":false},"security_clearance":false,"languages":[]},"benefits":[],"hiring_locations":[],"hiring_excludes":[],"relocation_offered":false,"industries":["Cloud Platforms (IaaS & PaaS)","Office & Productivity Software","Gaming Consoles","Developer Tools"],"lifecycle":[{"event":"open","at":"2026-11-18T20:26:38Z"}],"visa":[{"country":"US","licensed_sponsor":true,"adverts_offering_sponsorship_24m":0,"sanctioned_now":false,"evidence":"H-1B filings in 12 months: 6,135 · green card filings: 1,634","filings_12m":6135,"filings_prev_12m":9275,"green_card_filings_12m":1634,"median_offered_wage_usd":170139,"route":null,"cap_exempt":false,"checked_at":"2026-11-18T07:32:00+00:00","sources":["US Department of Labor: LCA disclosure data (H-1B, H-1B1, E-3)","US Department of Labor: PERM disclosure data (green cards)"],"filings_for_role_12m":4594}],"liveness":{"score":86,"band":"hot","label":"Hiring now","p_open":1,"p_active":0.86,"p_room":1,"age_days":1,"expected_fill_days":30,"reasons":["conf:2","win:early","comp:junior,brand"],"computed_at":"2026-11-18T22:55:04Z"},"pay":{"stated_usd_annual":161760,"is_top_pay":true},"html_url":"https://alion.io/job/microsoft-research-intern-self-improving-secure-ai-agents-microsoft-research","json_url":"https://alion.io/job/microsoft-research-intern-self-improving-secure-ai-agents-microsoft-research.json","meta":{"generated_at":"2026-11-18T22:55:04Z","cache_seconds":300,"methodology":"https://alion.io/methodology","terms":"https://alion.io/terms","contact":"https://alion.io/contact","api":"https://alion.io/developers","about":"Alion is a live layer of people, companies and AI agents: who they are, whether they are real and active right now, what they do and how to work with them, readable by people and by agents and paid per call.","catalog":"https://alion.io/catalog.json","usage":{"tier":"crawler_verified","counted_by":"address","units_charged":1,"used_today":3844,"day_limit":null,"remaining_today":null,"minute_limit":300,"resets_at":"2026-11-19T00:00:00Z"}},"offers":[{"id":"company.slices","title":"One company in depth, by slice","status":"live","price":{"credits":0.1,"usd":0.01,"plus_per_slice":{"credits":0.05,"usd":0.005}},"unit":"per company, plus each slice with data","note":"the employer in depth","call":{"mcp_tool":"get_company","arguments":{"id":18},"rest":"https://alion.io/mcp/rest/get_company?id=18"},"human":"https://alion.io/catalog?offer=company.slices&for=job%2Fmicrosoft-research-intern-self-improving-secure-ai-agents-microsoft-research"},{"id":"market.stats","title":"A market slice: pay, demand and time to fill","status":"free","price":{"credits":0,"usd":0},"unit":"per slice","note":"pay, demand and time to fill for this role and place","call":{"mcp_tool":"market_stats"},"human":"https://alion.io/catalog?offer=market.stats&for=job%2Fmicrosoft-research-intern-self-improving-secure-ai-agents-microsoft-research"},{"id":"job.search","title":"Open jobs by role, technology, place, pay and visa","status":"live","price":{"credits":0.1,"usd":0.01},"unit":"per posting in a list","note":"similar open postings","call":{"mcp_tool":"search_jobs"},"human":"https://alion.io/catalog?offer=job.search&for=job%2Fmicrosoft-research-intern-self-improving-secure-ai-agents-microsoft-research"},{"id":"company.verify","title":"Is this company real and active right now","status":"pilot","price":null,"unit":"per company","request":{"url":"https://alion.io/catalog/request","method":"POST","body":"{\"offer\": \"company.verify\", \"for\": \"job/microsoft-research-intern-self-improving-secure-ai-agents-microsoft-research\", \"note\": \"what you need it for\"}"},"human":"https://alion.io/catalog?offer=company.verify&for=job%2Fmicrosoft-research-intern-self-improving-secure-ai-agents-microsoft-research"}],"agent":{"mcp":"https://alion.io/mcp/jobs","mcp_open":"https://alion.io/mcp/open","next":[{"tool":"get_job","arguments":{"id":4559601}},{"tool":"search_jobs","arguments":{"company":"microsoft.com","detail":"compact"}},{"tool":"match_people","arguments":{"job_id":4559601},"note":"people open to offers who fit this vacancy"}],"note":"The same data through MCP tools: https://alion.io/mcp/open needs no sign-in (compact searches, postings, people); https://alion.io/mcp with an Alion account has everything. Docs: https://alion.io/developers"}}