The International Programme on AI Evaluation: Capabilities & Safety runs from 1 February to 25 June 2027, mostly online, and selects 40 fully funded participants from around the world. Applications close on 11 December 2026. At Youth Works Hub, we read the organiser's official application page and set out every date, condition and cost below. This is an academic programme run by ValgrAI: it is not part of Erasmus+ or the European Solidarity Corps (ESC), and it is listed in the Others category.
What the Second Edition of the Programme Is
🧠 Focus: building a new science of AI evaluation, with rigorous frameworks for evaluating frontier AI systems
🏛️ Organiser: ValgrAI (Valencian Graduate School and Research Network of Artificial Intelligence), with funding from Coefficient Giving and administrative support from BERI
🗓️ Dates: 1 February to 25 June 2027
💻 Format: mostly online, with live sessions every week
📍 Possible in-person week: Valencia, Spain, in late May or early June 2027, subject to funding
🎓 Credential: Expert Diploma certified by ValgrAI, with a grade transcript
────────────────────────────────────────
🌍 All opportunities in Spain →
────────────────────────────────────────
Why the Organisers Say AI Evaluation Matters
The page states that the world does not fully understand AI yet, and that the programme is building that understanding by turning scattered insights into a shared discipline. Its stated aims are to create the standards and language that governments, labs and societies can rely on, and to make AI accountable by testing and measuring it.
🔬 Pioneering a new science: AI evaluation as a discipline of its own
🤝 Shared language: a common vocabulary that bridges research, governance and policy
🛡️ Accountability: testing and measuring AI so that societies can trust how it is deployed
What You Gain from the Programme
📐 Rigorous frameworks: methods for evaluating frontier AI systems
🗣️ A shared language: a way to connect research, governance and policy
🌐 A global community: peers, mentors and faculty from leading institutes
🧪 Hands-on projects: evaluating real models and presenting capstone projects to experts
🚪 A direct path: towards opportunities in AI Safety Institutes, policy bodies and research labs
Who the Programme Is For
The page says this is for people who want to shape the safe future of AI: those who see the gaps and the unknowns and choose to lean in. It adds that the programme is not for anyone who wants a quick credential, and is for those who want to join the first generation of AI evaluators.
🎓 Graduate and PhD students
🔬 Early-career researchers
💼 Professionals with expertise in AI and machine learning, from academic training or professional experience
🧭 Exceptional candidates from policy, governance or social sciences
🌍 Applicants from around the world: the page names no nationality restriction
Last year's cohort included ML engineers, PhD students, a lawyer, people from AI safety institutes, cognitive scientists and policy analysts. The organisers say this diversity is intentional.
Selection Criteria
The page lists what the programme team looks for in candidates:
🎯 Clear motivation: a specific plan for where you are going and where you would like to have an impact in AI evaluation and safety
🧠 Expertise: a background in AI and machine learning
🌐 Diversity of perspective: across disciplines, geographies and professional backgrounds
📈 Potential to contribute: to research, policy or practice in AI evaluation after the programme
🔍 Curiosity: about how other fields can add to AI evaluation, and about developing a shared language
📅 Commitment: you have made space in your calendar and are ready to learn and actively contribute
🧮 Some technical literacy: for example reading a paper and spotting the strength of the method, understanding introductory statistics, calling a model through an API, running a notebook and interpreting a benchmark score
Technical literacy is not absolutely required because it is part of the course, but previous experience helps, especially for non-technical backgrounds. If you are working on your application statement, our guide on how to write a scholarship motivation letter explains structure and length.
Selection Process
📝 Application form: submit your CV and motivation statements
📋 Shortlisting: the programme team reviews all applications using transparent and equitable criteria; those who clear the bar go to the academic directors for a technical review and manual scoring
🗣️ Interviews: shortlisted candidates are invited to a 25-minute conversation; last year about 100 people were interviewed
✅ Final selection: offers go to 40 participants, including a few waitlisted; last year some waitlisted candidates joined when others declined
Programme Structure and Workload
The programme is organised into three parts, with one module a week and a buffer week after each part. The full diploma needs all three parts plus the capstone.
1️⃣ Part I, Foundations: Modules 1 to 4, February
2️⃣ Part II, Evaluation Methods: Modules 5 to 8, March to mid-April
3️⃣ Part III, Impact: Modules 9 to 12, late April to mid-May
🏁 Capstone: after the modules end, a few weeks are set aside to focus exclusively on the capstone
A module week has readings and short recordings before the week, two live sessions that mix teaching, discussion and exercises, one graded activity due within two weeks, and a Slack community throughout.
⏱️ Weekly time: 12 to 16 hours, made up of 2 to 3 hours of readings and recordings, about 3 hours of live sessions, 3 to 4 hours on the module activity and, from March, 4 to 6 hours on the capstone
🕘 Live session slots: the application asks which of four slots you could make most weeks: Tuesday or Thursday, around 9:00 or around 17:00 Central European Time; the team picks the most popular so that everyone can attend at least one live session each week
📅 Attendance: you must attend at least one live session each week; for sessions you miss, you watch the recording and do the exercise; four or more missed weeks overall means withdrawal from the programme
🔁 Buffer weeks: built in so you can rest or catch up
The page warns that if your honest answer is "I can do 8 hours a week", that is not enough time. It also says to tell the team early if you fall behind, but a participant who stops attending and stops handing in work without talking to the team may be asked whether this is still the right time.
Assessment and Credential
Everything is marked out of 10 and you need an average of 5 to complete.
📊 Module activities: about 50% of the final grade
🧪 Hands-on modules: 15%
🎓 Capstone: 35%
📜 Diploma: Expert Diploma certified by ValgrAI, with a grade transcript; the workload is equivalent to 15 ECTS but is not transferable as university credits
⏭️ Skipping modules: Modules 2 and 3 (architectures, metrics and experimental methodology) can be skipped if you can prove your knowledge, approved case by case
Costs, Funding and Visa Information
🎁 Tuition: you do not have to pay; tuition is covered by a scholarship funded by Coefficient Giving
🚫 Stipend: there is no stipend, so no money for living costs or lost income
🛂 Visa: you may need a Spanish visa depending on your passport; if you do, the organisers provide an official invitation letter and guidance, and you pay the visa fee and travel insurance
⏳ Timing: Schengen applications can take 4 to 8 weeks, so tell the team early; about a third of last year's cohort needed a visa and two of them were not able to get one
🏨 In-person week: the page does not describe costs for the possible Valencia week; confirmation depends on funding
The In-Person Capstone Week in Valencia
The organisers hope to hold one week in Valencia, Spain, in late May or early June 2027, and are working on the funding. Three things can happen: everyone goes; some teams go, chosen by the April capstone checkpoint grade, while the rest present online the same week; or it all runs online. The team says it will tell participants as soon as it knows.
Important Dates
📣 Applications opened: 4 November 2026
⏳ Final deadline: 11 December 2026, 23:59 Anywhere on Earth (AoE)
📬 Final decisions released: 15 January 2027
✍️ Deadline to accept offers: 22 January 2027
▶️ Programme starts: 1 February 2027 (online)
🏛️ In-person capstone week in Valencia: late May or early June 2027, subject to funding
🏁 Programme ends: 25 June 2027
How to Apply
Applicants complete the online application form and submit a CV and motivation statements. The form also asks which of the four weekly live-session slots you could attend most weeks.
Frequently Asked Questions
Is the International Programme on AI Evaluation free?
Tuition is covered by a scholarship funded by Coefficient Giving, so participants do not pay it. There is no stipend, and participants who need a Spanish visa pay the visa fee and travel insurance themselves.
Who can apply to the 2027 programme?
Graduate and PhD students, early-career researchers and professionals with expertise in AI and machine learning can apply, as can exceptional candidates from policy, governance or social sciences. The page says the organisers are mostly interested in your motivation to work on AI evaluation and safety.
Do I need a technical background?
You do not need to be an engineer, but some technical literacy helps, such as reading a paper, understanding introductory statistics, calling a model through an API and interpreting a benchmark score. It is part of the course, so it is not absolutely required.
How many hours a week does the programme take?
Plan for 12 to 16 hours a week, including readings, about 3 hours of live sessions, module activities and, from March, capstone work. The organisers say 8 hours a week is not enough.
Do I have to attend the live sessions?
Yes, you must attend at least one live session every week. If you miss one, you watch the recording and do the exercise, and four or more missed weeks overall means you are withdrawn from the programme.
Will I get university credits?
You receive an Expert Diploma certified by ValgrAI with a grade transcript. The workload is equivalent to 15 ECTS, but it is not transferable as university credits.
Is the Valencia week guaranteed?
No. The in-person capstone week in late May or early June 2027 is subject to funding, and the programme may run entirely online.
Is this an Erasmus+ or ESC project?
No. It is a ValgrAI programme and is not run under Erasmus+ or the European Solidarity Corps.
Deadline
⏳ 11 December 2026
Links
🔗 Application form: https://ai-evaluation.org/apply
🔗 Programme website: https://ai-evaluation.org/
🔗 ValgrAI programme page: https://valgrai.eu/ai-evaluation-programme/
✉️ Contact: info@valgrai.eu
