Imperial Schmidt AI in Science Global Faculty Fellowship 2027: £48,000 Tax-Free Stipend in London, Deadline 24 February 2027

Imperial College London main entrance on Exhibition Road

A year in London, a tax-free stipend of £48,000, and a second year funded back at your home university. That is what Imperial College London is offering early and mid-career scientists who want artificial intelligence to change their field, not just speed it up.

The Eric and Wendy Schmidt AI in Science Global Faculty Fellowship, supported by Schmidt Sciences, opened applications for its 2027 cohort on 3 November 2026. The deadline is 24 February 2027. It is a genuine two-year fellowship with a clear purpose: take a talented faculty member from a partner institution, give them a year inside Imperial's AI community, then fund them to carry that work home.

There is one catch worth stating up front, because it decides whether you should read further. This fellowship is only open to faculty at six partner institutions. If you are not at one of them, this is not your call. If you are, it is one of the best-funded AI fellowships open right now.

At a Glance

DetailWhat it is
FellowshipEric and Wendy Schmidt AI in Science Global Faculty Fellowship 2027
HostImperial College London, I-X Centre for AI in Science (Digital Foundry)
FunderSchmidt Sciences
DurationTwo years: one year in London, one year at your home institution
Stipend£48,000 tax-free for the year in London
Extra funding£30,000 flexi-fund per fellow for travel, conferences and consumables
Start date1 September 2027
Deadline24 February 2027 (check your institution for an internal date)
Application feeNone mentioned on the official page
Who can applyEarly and mid-career faculty at six partner institutions (see below)

What the Fellowship Includes

The first year is spent fully funded at Imperial, based at Digital Foundry, the college's interdisciplinary AI initiative. Fellows work alongside Imperial's AI faculty and around 25 AI in Science Fellows, with access to GPU compute and dedicated support staff for AI training and career development. Before the year starts, Imperial runs a training needs assessment and builds a bespoke training plan with each fellow, so the year is shaped around the skills that person actually needs.

The money is generous and it is structured carefully. The £48,000 stipend for the London year is tax-free. If you bring a partner or dependent children, Imperial pays for their flights, the dependent visa, and the immigration health surcharge, and offers additional financial assistance for dependent children. The £30,000 flexi-fund covers travel, conferences, and consumables, managed by Imperial in year one and by the home institution in year two.

Year two is the part most fellowships forget. Funding goes directly to the partner home institution to pay for a partial buyout of the fellow's time, so the fellow can keep focusing on the research instead of being pulled straight back into a full teaching load. The home institution hires a teaching fellow or equivalent to cover the gap. There are also bidirectional visits between Imperial and the partners, cohort events, seminars and socials, connections to the Women in AI network run from Digital Foundry, and an annual fellow-led conference with a prize for the best AI in Science work from global researchers.

After the fellowship ends, the support does not stop. Fellows get post-acceleration career support, including routes to seed funding and international grants, and membership of the international Schmidt Sciences AI in Science Fellow network, with links to partner programmes in Oxford and Cambridge and to Imperial's global hubs in Accra, Bangalore, and Singapore.

Who Can Apply

Eligibility is narrow on institution and broad on discipline, and that combination is deliberate.

The six partner institutions are the Indian Institute of Science, the National Centre for Biological Sciences in India, the International Centre for Theoretical Sciences in India, the University of Ghana, the University of Sao Paulo, and Universiti Teknologi Malaysia. Applicants must be faculty at one of these institutions. On average, one fellow per year is selected from each partner, except the University of Sao Paulo, which sends around three.

Within those institutions, the door is wide open. Candidates can come from any department, and the programme says plainly that most fellows will not come from Computing departments and will not hold degrees in computer science. What matters is that you are early or mid-career and that your work sits inside the AI in Science remit.

The remit is worth reading carefully, because there is no flexibility about it. AI is interpreted very broadly, including Bayesian inference and robotics. Science covers any typical topic in natural science and engineering. Epidemiology, biology, and basic science in biomedicine are included. Clinical medical themes are not covered, including conventional medical imaging. The social sciences and humanities are not covered either, apart from epidemiology. And the fellowship is not for research into generic AI with general application. You must be aiming to substantially advance one particular area of science, either as an AI researcher tackling a specific piece of science or as a scientist using AI to transform your area.

Example topics the programme names include Bayesian optimisation for molecular or materials design, machine learning for single-cell data, physics-based machine learning for turbine design, satellite imaging to predict disease risk, and astrostatistics. If your proposal reads like that list, you are in the right place. If it reads like a general AI methods paper, you are not.

One more thing the programme states clearly: a deep knowledge of AI is not a precondition. What it asks for is an appreciation of the need for AI and a willingness to learn it. Applications from women are particularly encouraged, and the programme aims for a gender-balanced cohort.

How Fellows Are Selected

Selection runs in two stages. Each partner institution forms its own internal review committee, which reviews and ranks the proposals from its own candidates. The committee looks at three things: your track record and your potential to grow as a research leader through AI, including willingness to take risks, creativity, and curiosity; your research statement, meaning the vision, the approach, the feasibility, and the potential for groundbreaking work and a sustained research portfolio afterwards; and your potential to acquire strong AI skills and the long-term impact of that learning.

Each institution then submits a ranked list of its top three candidates to Imperial's selection committee. The University of Sao Paulo submits its top six. Imperial conducts interviews as needed and makes the final selection.

How to Apply

Applications go to your home institution, not directly to Imperial, as a single file. The file has seven parts, and the page gives exact length limits for each, so follow them.

First, your CV including publications. Second, a publication elaboration of one page or less, outlining the research contribution of up to three of your papers for a general scientific reader. Third, a research proposal summary of one page or less, suitable for a general audience, including the project title. The programme says particular attention is paid to this summary, so it carries real weight. Fourth, a research proposal of three pages or less explaining why and how the work could be transformative for one area of science, built around background, a small number of hypotheses or aims, and work packages.

Fifth, a training plan of half a page or less, naming the skills you need to acquire for the research to succeed. Sixth, a quarter-page outline of how your proposal fits the AI in Science remit. Seventh, a quarter-page kindness statement, your view on the need for kindness among scientists.

On top of the file, you need two letters of support: one from your academic host at Imperial, using the template provided, and one from your Head of Department at home. Here is the step people get wrong. You must identify and contact an Imperial host before you submit, and applications without an identified host will not be considered. The host must be a permanent member of Imperial's academic or research staff, and hosts are not obliged to accept. If your first choice does not reply, contact others. Start this step early, because it is the slowest part of the whole application.

And one rule that matters for readers of this blog in particular: the official page states, in bold, "Please do not use LLMs in any part of your application, as doing so will make you ineligible." Do not use AI to write any part of this application. Write it yourself.

Important Dates

DateWhat happens
3 November 2026Applications opened
20 November 2026Information webinar
24 February 2027Application deadline (check your institution for an internal date)
Week of 8 March 2027Application review and interviews at Imperial
1 April 2027Fellows announced
1 September 2027Fellowship begins

Note the warning in the timeline: your home institution may set its own internal date before 24 February, because it needs time to run its internal review and ranking. If you are at a partner institution, ask your department now. Waiting until February to discover an internal January deadline would be an expensive mistake.

Should You Apply?

Apply if you are faculty at one of the six partners, you have a specific scientific problem, and you can explain in one page why AI changes that problem. The funding is real, the London year is fully covered, and the second-year buyout is the kind of detail that shows the programme was designed by people who understand academic life.

Be honest with yourself about the fit. If your work is clinical medicine, this is not your fellowship. If your research is general AI methods with no single scientific target, this is not your fellowship. And if you cannot name an Imperial host who works anywhere near your area, that is the first problem to solve, not the last.

For the right candidate, though, this is a strong call. A tax-free £48,000 stipend, a £30,000 flexi-fund, family support, GPU access, a bespoke training plan, and a funded second year at home add up to one of the most complete fellowship packages open to AI researchers right now.

If a permanent faculty post is what you are after rather than a fellowship, the University of Notre Dame tenure-track faculty job in responsible AI is open rank, with full consideration by 15 November and a final deadline of 15 December 2026. The CISPA tenure-track faculty post in trustworthy AI in Saarbrücken, Germany, is another permanent faculty route, with a deadline of 3 January 2027.

Frequently Asked Questions

Who is eligible for the Imperial Schmidt AI in Science Global Faculty Fellowship?

Early and mid-career faculty members at six partner institutions: the Indian Institute of Science, the National Centre for Biological Sciences, the International Centre for Theoretical Sciences, the University of Ghana, the University of Sao Paulo, and Universiti Teknologi Malaysia. Applicants can come from any department, and a computing background is not required.

What is the deadline for the 2027 cohort?

24 February 2027. Your home institution may set an earlier internal deadline, because it runs its own review and ranking first, so check with your department before planning around the February date.

How much funding does the fellowship provide?

A tax-free stipend of £48,000 for the year in London, plus a £30,000 flexi-fund per fellow for travel, conferences, and consumables. Family support covers partner and dependent flights, dependent visas, and the immigration health surcharge. The second year is funded at the home institution through a partial teaching buyout.

Is there an application fee?

No fee is mentioned anywhere on the official programme page. Applications are submitted as a single file to your home institution, not through a paid portal.

Do I need to be an AI expert to apply?

No. The programme states that a deep knowledge of AI is not a precondition, only an appreciation of the need for AI and a willingness to learn it. Most fellows are not from computing departments. You do need a specific scientific problem that AI can transform.

Can I use AI tools to write my application?

No. The official page states that applicants must not use large language models in any part of the application, and doing so makes you ineligible. Write every part yourself.

Apply on Official Website

Apply on the official Imperial College London programme page. All details in this post were checked against that page on 11 November 2026. If anything on the official page changes, the official page is the version that counts.

Sources

  • Imperial College London, I-X Centre for AI in Science, "Eric and Wendy Schmidt AI in Science Global Faculty Fellowships", applications opened 3 November 2026, deadline 24 February 2027: https://www.imperial.ac.uk/ix-ai-in-science/-eric-and-wendy-schmidt-ai-in-science-global-faculty-fellowships/

By Kashan Ajmeri, 11 November 2026. You may also like the Wallenberg-NTU Presidential Postdoctoral Fellowship 2027, the GovAI Summer Fellowship 2027 in London, and the Impact AI Fellowship 2027 in Berlin, three more funded AI fellowships on this blog.

Related articles

AI Singapore PhD Fellowship 2027: Up to S$6,700 a Month, Deadline 31 January 2027

Post a Comment

Previous Post Next Post