Section 1 is for most paid positions, offered through an advertised call. Section 2 covers ordinary thesis supervision. Section 3 explains what we do. Section 4 covers the most common questions about joining the group without an open call, and about other forms of academic collaboration. Section 5 covers employment conditions, and Section 6 covers other frequently asked questions.
Choose your situation:
Open vacancies:
If no positions are listed above, none are currently open.
We recruit primarily through advertised calls rather than unsolicited applications. Please follow the instructions in the advertisement and submit your application through the specified recruitment system.
A separate email is unnecessary unless the advertisement asks applicants to contact a potential supervisor in advance. As a general rule, we do not evaluate an applicant's fit or discuss funding by email before the call closes, except in truly exceptional circumstances; all applications are considered together after the deadline (see also FAQ 6.1). Please submit your application; we will be in touch after the call closes.
Some recurring recruitment routes are listed below. Our group's participation and the available positions and research topics vary between calls and from year to year.
This section covers ordinary BSc and MSc thesis supervision undertaken as part of a University of Helsinki degree. It does not cover paid positions, internships or doctoral supervision, which are covered in Sections 1 and 4.
I supervise BSc theses only through the University of Helsinki
I can serve as the primary supervisor of an MSc thesis only if the student has active study rights in one of the following University of Helsinki degree programmes:
This is an academic supervision arrangement undertaken as part of the student’s degree and does not normally include employment or a salary. Students seeking this ordinary form of MSc thesis supervision do not need to satisfy the exceptional-candidate criteria for paid positions. They must nevertheless have sufficient relevant academic preparation, propose a feasible topic within my areas of expertise, and take responsibility for planning and completing the thesis in accordance with their degree programme’s procedures.
The availability of supervision depends, for example, on:
Typically, the student proposes a topic for their thesis. A thesis may also be
The thesis topic must fall within an area in which I can provide appropriate academic supervision. It should normally concern machine learning, artificial intelligence, exploratory data analysis or a closely related methodological area.
Before supervision is formally agreed, I normally ask the student to complete my
You may contact me if you meet the criteria above and are prepared to follow the process described in the applicable supervision agreement. If you are an eligible University of Helsinki student seeking ordinary MSc thesis supervision and have not received a reply within a few working days, please resend your message.
In exceptional circumstances, I may co-supervise an MSc thesis at another university or in another University of Helsinki degree programme. Such an arrangement must be proposed by the thesis’s primary supervisor, normally a colleague at a Finnish university. I do not consider direct co-supervision requests from students at other universities.
This is unpaid academic supervision as part of your degree; for a paid MSc thesis project or internship, see Section 4.1.2 instead.
Machine learning and artificial intelligence are used extensively in the sciences. When modelling physical systems, the understandability and statistical robustness of models may be more important than predictive accuracy alone.
We do fundamental machine learning research on probabilistic models, the statistics of machine learning, and explainable and interpretable AI (XAI). One of our main application motivations is the natural world — measured and simulated phenomena — where our objective is to find ways for scientists and others to understand data and the underlying processes, and to build better models of nature. Please see
We work in a multidisciplinary team of computer scientists and atmospheric scientists. The group is part of the
VILMA aims to model atmospheric molecular-level processes efficiently and to understand their underlying mechanisms and causal connections. It combines first-principles quantum-chemical and other simulations with probabilistic machine learning and artificial intelligence models, interactive visualisation and exploratory data analysis.
For more general VILMA positions that are not specifically focused on artificial intelligence, please see the
Most of our work concerns methodological and theoretical research in machine learning and artificial intelligence. Although the methods are often applied to atmospheric and other natural-science problems, they are not specific to those domains.
People joining the group should contribute to the machine learning and artificial intelligence research agenda described above. A suitable project should normally contain an original methodological research question. Merely applying an existing machine learning model to a new dataset is generally not sufficient. For example, applying an existing model to a new atmospheric dataset is not usually sufficient on its own; developing, analysing, or validating a new method that addresses a general limitation may constitute an appropriate methodological contribution.
Topics are not automatically a good fit merely because they involve AI, machine learning or data. For example, cybersecurity, network traffic, software engineering, generic applications of large language models, social analysis of AI, recommender systems or reinforcement-learning applications may be outside our current focus unless the proposal contains a strong methodological question directly connected to our ongoing work.
Before contacting me, read the description of the group’s research and some of our recent publications. Your message should identify the specific part of our work to which your proposed project relates.
Candidates for research-group positions and exceptional expressions of interest should have a strong foundation in machine learning and the relevant mathematics, together with suitable programming skills.
We consider candidates with backgrounds including computer science, mathematics, atmospheric science, physics, and chemistry.
Interest in atmospheric science and the other natural-science topics described above is an advantage but not a requirement. Prior knowledge of atmospheric processes is optional. A broader interest in the natural sciences is advantageous.
Doctoral researchers must normally have completed an applicable master’s degree before starting employment; there is no direct route from a bachelor’s degree to doctoral studies in the Finnish setting. Doctoral researchers in our field are commonly employed in salaried positions while completing their degrees. Doctoral research normally takes three to four years, although the duration of an employment contract is agreed separately.
Postdoctoral researchers should have completed a PhD before starting employment. Postdoctoral contracts are normally between one and four years, with the duration agreed individually.
We welcome qualified doctoral and postdoctoral candidates from Finland and abroad. The study-right restrictions described above apply only to BSc and MSc thesis supervision and internships, not to advertised doctoral or postdoctoral positions.
This section concerns unsolicited expressions of interest in paid positions, doctoral supervision, postdoctoral or project-researcher opportunities, scientific collaboration, and research visits. We remain interested in hearing from exceptionally strong postdoctoral researchers, doctoral researchers and MSc students whose expertise and proposed work closely match our current research. In rare cases, funding or another suitable research-group arrangement may become possible even when there is no open call.
This should not be understood as an open position or a general invitation to request doctoral supervision, employment, scientific collaboration, co-authorship or a research visit. Both demonstrated excellence and a close research fit are required. Please read the relevant sections of this page carefully before contacting me.
If your background, proposed work, and intended arrangement closely match the criteria below, we welcome a concise, well-prepared message.
However, we receive many requests for employment, doctoral supervision, research visits and meetings, and in practice the great majority are not a realistic match for either side. The criteria below are meant to help you work that out before writing rather than after — a better use of your time than a message that, realistically, is unlikely to lead anywhere.
These situations fall into three different kinds. In Section 4.1, I would be your employer or host, and I assess you as a candidate against the criteria below. In Section 4.2, someone else is responsible for your day-to-day supervision, and what matters instead is the fit of the work and the credibility of that arrangement. In Section 4.3, the arrangement is normal peer-level academic collaboration rather than supervision at all, assessed on a different basis again.
A note on Finnish doctoral study: The Finnish system differs from models where a department admits a student and assigns a supervisor afterwards. At the University of Helsinki, you typically identify a supervisor and agree on a supervisory arrangement before applying for
In each of the following, I would be directly employing or hosting you: a paid MSc thesis or internship (4.1.2), doctoral research (4.1.3), or a postdoctoral or project-researcher position (4.1.4). I assess you against the exceptional-candidate criteria just below.
These criteria govern the arrangements in this section — a paid MSc thesis or internship, a doctoral position, and a postdoctoral or project-researcher position. They do not apply to ordinary BSc or MSc thesis supervision for eligible University of Helsinki students (Section 2), nor to the arrangements in Sections 4.2 or 4.3, where someone else is primarily responsible for your supervision, or the arrangement is peer-level collaboration rather than supervision at all, and a different standard applies in each case.
“Exceptional” does not mean simply motivated, hard-working or interested in the group. These qualities are expected of all serious applicants.
Evidence of exceptional preparation may include a combination of:
I assess expressions of interest using three candidate-related criteria, all of which must be satisfied:
In addition, a suitable project, sufficient supervision capacity and appropriate funding must be available. Even an exceptional candidate cannot be hired if these practical conditions are not met.
The main channel for a paid MSc position is the annual Faculty of Science summer jobs call (Section 1); this section covers the underlying criteria, which also apply if you approach us directly outside that call.
If you are looking for ordinary (unpaid) thesis supervision as part of your degree, see Section 2.2 instead.
Paid MSc thesis projects and internships are separate from ordinary degree supervision. They involve research-grade problems closely aligned with the group’s scientific interests and are available only when suitable funding, supervision capacity and a suitable project exist. Most of our paid MSc students have started as summer interns, and most of our doctoral researchers have previously worked with us as interns or MSc students.
For paid MSc thesis and internship positions, we primarily consider students who:
Students who have performed well in my courses are among those who may be considered. We primarily seek students who may be interested in continuing to doctoral research in the group after completing their MSc degree.
When funding and a suitable project are available, possible paid arrangements may include:
An internship may later lead to an MSc thesis project, and an MSc thesis project may later lead to doctoral research. Any continuation requires mutual agreement, satisfactory progress, a suitable topic, available supervision and, where employment is involved, separate funding; it is never automatic.
The MSc thesis restrictions in Section 2.2 apply to paid MSc thesis projects. For internships, I consider only students with active study rights at the University of Helsinki. I do not host interns who are enrolled only at another university.
There are three possible ways to become a doctoral researcher in the group. The usual and most reliable route is an advertised, project-funded doctoral position (Section 1). The second route is an approach without an open call: in rare, exceptional cases, an unsolicited approach from an outstanding candidate may lead me to arrange project funding even without an open call, judged against the same criteria as an advertised call (Section 4.1.1). If you have a way to fund your own position — for example, a personal fellowship or a grant you would apply for yourself — this can also work in principle without an open call, provided the arrangement otherwise meets the exceptional-candidate criteria and I have the necessary supervision capacity. The third but in practice most unrealistic route is to apply independently for a
If you wish to propose a supervisory arrangement under any of these routes, I will assess your request according to the standards that I apply when assessing candidates for a paid doctoral researcher position in my group. Please follow the instructions under the call text (Section 1) or how to contact me (Section 4.4) and provide the requested application materials, including a motivation letter, CV and study transcript.
For an advertised position (Section 1) or an unsolicited request, simply apply as described above — there is normally no need to discuss the proposal with me in advance (Section 6.1). The exception is the University-funded doctoral researcher position, which requires a detailed, jointly developed research plan: establishing that arrangement and preparing a competitive proposal takes substantial time, typically months, so contacting me a month before the deadline is too late. I agree to act as supervisor for that call for a candidate I do not already know only in exceptional cases (Section 4.1.1).
After the research, supervision, employment and funding arrangements have been established via an advertised position, the researcher
Postdoctoral and project researcher positions will be advertised on this page (Section 1) and through the University’s recruitment channels when they become available.
If no position is advertised, we can consider only rare expressions of interest from researchers who have:
A substantive recommendation from a respected researcher who knows your work well is helpful, particularly if the researcher is someone I know professionally. It is especially helpful if such a colleague contacts me directly to explain the significance of your work and why they consider you a strong match for the group; requests supported or recommended by a researcher whom I know professionally are generally easier to evaluate than cold approaches from researchers whom I do not know. A recommendation does not replace the requirements above or guarantee consideration. To be clear, this is not a requirement — a strong cold application from someone I don't yet know is equally welcome; a recommendation simply gives me useful context faster.
A general overlap in terms such as artificial intelligence, machine learning, large language models, explainability, data mining or data science does not by itself establish a sufficiently close research fit.
Postdoctoral applicants must have completed a doctoral degree before starting employment. Postdoctoral contracts are normally between one and four years, with the precise duration agreed individually.
In each of the following, someone else is primarily responsible for supervising you day to day: acting as a coordinating academic (4.2.2) or a secondary or methodological co-supervisor (4.2.3). Peer-level academic collaboration, co-authorship and research visits — not supervisory arrangements at all — are covered separately in Section 4.3.
For a coordinating-academic or co-supervision arrangement (4.2.2, 4.2.3), what matters is whether your day-to-day supervision is genuinely and credibly handled elsewhere, whether the work has a real methodological fit with mine, and the competence and track record of your actual supervisor(s), since I am relying on their judgement as well as my own. In both cases, whoever serves as coordinating academic must also fit the thesis's main discipline, diagnosed via target publication venues as in Section 4.2.2 — so if your main research question is not itself ML/AI methods development, the coordinating academic should come from your own field even where I am the right secondary supervisor for the methodological side (Section 4.2.3); if it is ML/AI methods development, I can combine both roles. Collaboration, co-authorship and a research visit are handled differently (see Section 4.3).
If your actual research supervisor is based outside the University of Helsinki — for example at a company or research institute — I may in some cases serve as your coordinating academic without research supervision responsibilities. The coordinating academic oversees the doctoral process and handles the relevant University administrative requirements, while your external supervisors remain responsible for directing your day-to-day research and for proposing your pre-examiners and opponent (who formally evaluate the completed thesis before and during the public defence) at the appropriate stage of the doctoral process. For the University’s general instructions, see
This is different from asking me to supervise your research: the coordinating-academic role does not include day-to-day research supervision, but it still involves a real ongoing commitment — reviewing your progress, evaluating the thesis, and organising the public defence — so I assess each request individually. I need to trust not only that you are competent and motivated, but that your external supervisor can reliably cover the computer science and machine learning content, with enough allocated time to do so.
This arrangement may be suitable if all the following apply:
Meeting these conditions is necessary but not sufficient: I also need sufficient capacity for the ongoing commitment described above.
If all of these conditions apply, you and your proposed supervisor should contact me together, well before any application deadline. In addition to the general application materials described in Section 4.4, your motivation letter should specifically describe the proposed research questions, thesis topic, target publication venues, and the intended supervision arrangement between you, your external supervisor(s) and me. If I do not already know your external supervisor, please also ask them to send their CV and list of publications, so that I can assess their capacity to supervise the computer science and machine learning content.
This scenario still needs a coordinating academic who fits your thesis's main discipline, using the same criteria as Section 4.2.2: someone whose research area matches your main research question, diagnosed via your target publication venues. I can combine this coordinating-academic role with the secondary-supervisor role below when your main research question is itself ML/AI methods development; otherwise, the coordinating academic should be someone from your own field.
This is a genuine, active supervisory commitment from me, not only administrative oversight: I contribute directly to the methodological and machine-learning side of your research, typically as a secondary or methodological co-supervisor rather than your main supervisor. Unlike Section 4.2.2, this is the right route when your existing day-to-day supervision at a collaborating research group, company, research institute or other organisation does not already cover the methodological and machine-learning side of your work, so active input from me is needed rather than administrative oversight alone.
If a collaborating organisation — responsible for your funding and day-to-day supervision — would employ or host you, the natural first step is for you and your intended supervisor(s) there to talk the arrangement through together and reach a shared understanding of the roles; after that, I'd want the organisation to confirm the arrangement with me directly — this protects you as well, since a commitment on my part only makes sense once funding and supervision are actually settled. A request from a student alone, without clear funding and supervision arrangements in place, will not normally lead to a commitment. If the host organisation has no PhD-level supervision capacity of its own, this lighter co-supervisor form is not available — the arrangement would instead require the same commitment from me as if you were employed directly in my group. Such cases are rare.
These arrangements are not supervisory relationships at all: they are peer-level academic collaboration, co-authorship or a research visit, and I assess them on the basis set out below rather than the criteria in Sections 4.1 or 4.2.
In short, as a general rule, unsolicited requests for collaboration, co-authorship or a research visit are considered only when proposed or recommended by a researcher I already know professionally.
Contact about collaboration, co-authorship or a research visit should come from whoever is responsible for the work. If you are a student (BSc, MSc or PhD), that is your supervisor, not you alone, and a request without your supervisor's involvement and support will not normally be considered. If you are a PI or otherwise responsible for your own research — for example, a postdoctoral researcher — you should contact me directly yourself. The same academic standards apply in all cases. A trusted colleague can often give a realistic assessment of the person, the project and the feasibility of the arrangement, which makes it much easier for me to decide whether committing my time is justified. Such a colleague should contact me directly and explain the project, the proposed roles, and why my contribution or hosting the visit would be useful.
An unsolicited approach from someone with whom I have no prior professional relationship or natural common ground rarely leads to substantial collaboration or co-authorship — not because the work is unwelcome, but because a cold approach gives me no reliable way to judge fit, and building that from scratch takes time neither side usually has. If there is someone we both know, asking them for a reference or introduction is a far more effective route than writing cold.
Whether I can participate in any collaboration or research visit depends on the scientific fit, the contribution expected of me, my available time, and the feasibility of the arrangement.
I consider unsolicited requests for exchange or other short-term research visits only when all the following conditions are met:
A request to host a visit should include a specific research plan rather than a general willingness to work on topics to be identified later.
External funding, including Erasmus+ funding, removes the direct financial cost of a visit, but it does not remove the supervision, planning, and working time required of the host group (see also FAQ 6.5). I cannot host visits whose connection to our work is only at a broad level, such as machine learning, artificial intelligence, or data analysis.
Authorship requires a genuine and sufficiently substantial contribution to the research. It cannot be offered merely in return for reviewing a manuscript, providing general advice, helping to obtain publication or allowing the use of my name. Authorship and the expected contributions should normally be discussed before substantial joint work begins. In particular, I do not normally join an existing project, manuscript or software system unless there is a clear scientific contribution that I can make and sufficient time to make it.
A University of Helsinki affiliation cannot be added merely because I advise on or co-author a manuscript. Each author’s affiliation must accurately represent the institutional context in which their contribution to the work was made.
The instructions in this section apply to unsolicited expressions of interest in paid positions, doctoral supervision, postdoctoral or project researcher opportunities, and research visits. University of Helsinki students seeking ordinary BSc or MSc thesis supervision should instead follow Sections 2.1 and 2.2. Scientific collaboration and co-authorship requests should first be assessed against Section 4.3. If Section 4.3 indicates that contacting me may be appropriate, the relevant information below should be adapted to the proposed arrangement.
If, after reading this page, you meet the relevant criteria, email me (
The letter should explain:
Please attach:
A substantive recommendation from a respected researcher who knows your work is helpful, especially if that researcher knows me professionally and contacts me directly. A brief, substantive email is sufficient unless a formal letter is requested. A recommendation does not replace the required materials or guarantee consideration.
A study transcript may be unofficial, but it must:
You do not normally need to send copies of publications, a copy of your thesis, degree certificates, formal reference letters, portfolios, certificates from short courses, or other supplementary documents not requested above.
Please do not send generic cover letters that could equally be addressed to any machine learning group.
A suitable email subject is: Expression of interest: [doctoral researcher / paid MSc project / internship / postdoctoral researcher / visit] | [specific research topic]
We will contact you if your background and interests match a realistic opportunity. Because of the volume of enquiries, we cannot respond to most unsolicited messages — if you do not receive a reply, you should assume that I do not currently have a suitable position, supervision arrangement, collaboration or visiting opportunity for you (see Section 6.9 on following up).
All employed positions include a salary in accordance with the relevant University of Helsinki collective agreement. Employees also receive occupational healthcare and other University staff benefits. Social security, pension contributions and access to Finland’s national health insurance system are determined in accordance with the applicable legislation and conditions.
The University covers work-related travel expenses in accordance with its policies.
Further information is available about:
A meeting is never the first stage of the process; it happens only once you have explained in writing what it would be for, and that suggests a realistic fit. As a general rule, we will only evaluate and comment on your materials by email if we see a realistic fit and opportunity.
In advertised calls (Section 1), we review the materials after the closing of the call, not before (except in truly exceptional circumstances): it is also much harder to evaluate an application fairly before seeing the other applications it will be compared against. A strong candidate with a well-written application will stand out in that review, whatever was or wasn't discussed beforehand. For these reasons, while contacting us in advance is allowed, it is not necessary and usually not even helpful, unless the call text explicitly invites it.
Please do not use my meeting-booking link to request an initial discussion about an unsolicited application, possible supervision, scientific collaboration or a research visit.
Normally, no. Applicants for research positions should be able to identify a research direction based on the group’s published work and explain how their background prepares them to contribute to it.
Topics can be developed jointly after a student or researcher has been selected and a realistic working arrangement has been established.
Possibly, but funding alone is not sufficient. Research fit, qualifications, supervision capacity, employment or affiliation arrangements, and the conditions of the scholarship must all be suitable.
Please do not apply for a scholarship on the assumption that supervision will be agreed afterwards.
As a general rule, no. Agreeing to act as a supervisor is a substantive academic commitment, not merely an administrative requirement for an application.
If I would not realistically be able to participate in the supervision arrangement, I will not agree to be named as a supervisor solely to enable an application.
Only after a substantive research and hosting arrangement has been assessed and agreed, using the same visit criteria as Section 4.3.2 — external funding for a visitor's expenses does not by itself satisfy them. This also applies to requests for hosting, support or invitation letters for visiting fellows, exchange, mobility or similar programmes. We cannot provide provisional invitation letters for speculative applications.
Interdisciplinary backgrounds can be valuable, particularly backgrounds in mathematics and the natural sciences. However, research in the group requires substantial preparation in machine learning, statistics, mathematics and programming.
Note that introductory coursework, brief online training and small tutorial projects are not normally sufficient preparation for research in the group.
No. These are important questions, but individual professors do not admit students to Bachelor's or Master's programmes and cannot assess admission prospects through informal email — please consult the
A potential supervisor cannot promise admission, and admission to a programme does not by itself create funding, employment or a supervisory commitment.
For doctoral studies, consult the relevant doctoral programme’s instructions, eligibility requirements, application periods and supervision requirements. Contact me only if the available materials do not answer your question and your question specifically concerns a realistic research or supervision arrangement with me.
I cannot normally offer individual meetings to discuss general degree options, admissions procedures, scholarships or how to begin a career in machine learning.
Please use the University’s research group, faculty and doctoral programme pages to identify researchers whose recent work directly matches your proposed topic. Because of the number and breadth of enquiries, we cannot provide an individual referral for every request.
Please do not send repeated follow-ups about unsolicited expressions of interest. If I identify a sufficiently close match and a realistic opportunity, I will contact you. Apply separately to every advertised position that interests you: an earlier email does not constitute an application for a later call and will not necessarily be reconsidered automatically.
Please see Section 4.3.1. As explained there, I normally consider such arrangements only when they are proposed or recommended by a researcher whom I know professionally.