Here is the short answer most guides bury: you do not apply to Mila or the Vector Institute for a PhD. You apply to a partner university's graduate program, and separately you line up a professor at the institute to supervise you. Both things have to happen for you to end up doing AI research at either place. Miss the fact that they are two parallel tracks and you can easily waste a full admission cycle, which for a research degree means losing a year.
This walks through how admission actually works at Mila in Montreal and Vector in Toronto, the timeline and the December 1 gate at Mila, how to find and contact a supervisor, and whether the PhD is funded. It is informational, not legal, immigration, tax, or financial advice, so verify every date and rule against the official pages before you act.
How getting into Mila or Vector actually works
Both institutes are research organizations, not degree-granting universities. Your PhD is awarded by a university; the institute is where the research and the community live. That single distinction explains almost every piece of confusion around applying.
At Mila, students are admitted to a graduate program at one of the partner institutions and then work under a Mila professor. The partners, as of 2026, are Universite de Montreal (through the DIRO computer science department), McGill University, Polytechnique Montreal, and HEC Montreal. Mila runs an annual supervision request process where you ask a specific core academic member to take you on. That request is separate from your university application, and both have to be moving in parallel. Per the official Mila supervision request page, you are told to register with the professor's department and university even before you have a confirmed supervisor.
At Vector, the model is looser. There is no single "apply to Vector" degree form. You enrol in a graduate program at an Ontario university (University of Toronto and many others), and your connection to Vector comes through a faculty member who is a Vector researcher, plus programs like the Vector Faculty Affiliates program and the postgraduate affiliate track for PhD students already at an Ontario institution. So the practical route into Vector is: get admitted to a strong Ontario AI lab whose supervisor is affiliated with Vector.
The two-track timeline and the December 1 gate
Mila's calendar is the one people trip over. The supervision request window generally opens in mid-October and closes on December 1 for the following Fall intake. That deadline is the gate: if your supervision request is not in by then, you are effectively out for that cycle regardless of how strong your university application is.
The catch is that the two tracks resolve on different clocks. Supervision decisions from Mila professors typically come back in February or March, sometimes slipping into April. Meanwhile the partner university has its own application deadline and its own admissions committee. You cannot wait to hear from a professor before applying to the university, because by then the university deadline may have passed. Run both at once.
Vector has no equivalent single institute-wide date, since admission is driven by each Ontario university's graduate deadlines (many fall in December or January). The discipline is the same: target the deadline of the specific program, and reach the affiliated supervisor early.
How to find and contact a supervisor
The supervisor decision matters more than the institute brand. A funded, active lab in your subfield beats a famous name who has no room and no time. A few things that actually move the needle:
- Read before you write. Pick two or three professors whose recent papers you have genuinely read, not just skimmed the titles of. Generic mass emails get ignored.
- Check they are recruiting. Many post "I am taking students for Fall 20XX" notes on their lab page or social accounts. If they say the lab is full, believe them.
- Keep the first email short. One paragraph on who you are, one on why their specific work fits what you want to do, and a link to your CV plus any code or publications.
- Name the fit, not the flattery. Reference a concrete problem from their work you would want to push on.
Because a PhD is a multi-year commitment to one advisor, this is worth the hours. If you are weighing Canada against other destinations first, our companion piece on AI PhD stipends by country is a useful reality check, and the full the AI Relocation Guide lets you compare all 21 countries on pay, visas, and years to permanent residency.
What to do this week
- Decide the subfield first. Write one sentence describing the research you want to do. It filters the supervisor list fast.
- Build a shortlist of 5 to 8 supervisors across Mila and Vector labs, noting for each whether they are recruiting and which university they sit under.
- Check the two deadlines for each. The Mila supervision window (mid-Oct to Dec 1) and the partner university's own graduate application deadline. Put both in a calendar.
- Draft the supervision emails now so you can send the moment the window opens, not scramble in late November.
- Start the university application in parallel, including transcripts, references, and English test scores, without waiting to hear from any professor.
Funding: is an AI PhD in Canada salaried?
Generally, yes, a research PhD at Mila or a Vector-affiliated lab is funded rather than something you pay for. Funding usually comes as a package: a stipend from your supervisor's grants, university teaching or research assistantships, and external awards on top. The most visible top-up in Ontario is the Vector Scholarship in Artificial Intelligence, a merit award (roughly 17,500 CAD as of 2026) aimed at students in Vector-recognized master's programs, which can also be a stepping stone toward a PhD.
Exact stipend amounts vary by supervisor, university, and year, so treat any single figure as directional and confirm it in your offer letter. After the PhD, the post-study work rules matter as much as the stipend; our guide to the PGWP for AI and CS graduates covers what comes next.
The honest takeaway
Mila suits you if you want a large, concentrated AI research community with a clear, calendar-driven process and are comfortable applying in French-adjacent Montreal (research is in English; daily life often is not). Vector suits you if you want the breadth of Toronto's universities and industry ties, and you are happy to route in through a specific Ontario program and supervisor rather than one central door. In both cases the real admission decision is made by a supervisor and a university, not by the institute's name on the building.
Rule of thumb: pick the supervisor and the deadline first, the institute logo second, and always run the supervision request and the university application at the same time.
