Should You Do a Masters or PhD After UofA CS?¶
This is a question most CS students ask themselves at some point, and the answer is genuinely different for different people. This guide gives you the real picture — not the admissions brochure version.
MSc vs PhD: What's the Actual Difference?¶
Masters (MSc): Typically 2 years. You take some courses, complete a thesis (or in some programs, a project report), and demonstrate that you can conduct independent research at a graduate level. You become an expert in a narrow area. At UofA CS, it's almost always thesis-based for research-focused students.
PhD: 4–6 years (sometimes longer). You produce original knowledge — something that did not exist before you worked on it. You defend a dissertation to a committee of experts. You become the expert on a very specific problem. It's a long, often non-linear path that involves a lot of uncertainty.
These are genuinely different experiences and serve different purposes. Don't conflate them.
Is an MSc Worth It for Industry?¶
Honest answer: it depends on what you want to do.
For most software engineering roles — backend, full-stack, mobile, DevOps, infra — an MSc provides marginal value over a strong BSc with good internship experience. Industry doesn't pay you for credentials; it pays you for what you can ship.
However, there are specific domains where an MSc genuinely moves the needle:
Machine Learning Engineering: Major tech companies (Google, Meta, DeepMind, etc.) increasingly expect MSc or equivalent research experience for ML roles. This has become more true as the field matures. If you want to work on the actual model training, architecture research, or applied ML infrastructure at a top company, an MSc is often the floor.
Research Engineering at Labs: Roles at OpenAI, Anthropic, DeepMind, Google Brain (now Google DeepMind) often explicitly require or strongly prefer graduate degrees.
Transitioning to Management Later: Some senior engineers and engineering managers find that grad school develops the communication, project scoping, and independent thinking skills that accelerate leadership trajectories. This is less about the credential and more about the experience.
The math you should actually do:
A BSc grad in CS at a decent company in Canada/US makes \(80K–\)120K CAD to start. An MSc takes 2 years. Even if an MSc bumps your starting salary by $15–25K (which it might, in ML-focused roles), you've lost 2 years of compounding salary and experience. Over a 10-year horizon, you likely break even. The calculation changes if you're at a top company in a specialized field, or if you care about the experience of research beyond just the salary.
An MSc is probably worth it if: you genuinely want to work in ML/AI research-adjacent roles, you want to stay in academia as an option, or you love the work of research enough that you'd value the experience independent of the salary delta.
An MSc is probably not worth it if: you want to be a software engineer at a tech company and don't have a specific domain that requires it. Go get experience instead.
Is a PhD Worth It for Industry?¶
Almost always no, unless your goal is research itself.
A PhD makes sense if: - You want to be a faculty member at a university - You want to work at a research lab (DeepMind, MSR, AI2, FAIR, etc.) in a research scientist capacity - You genuinely love the process of pushing knowledge forward and would do it regardless of the career outcome
A PhD is not a good choice if: - You don't know what you want to do and are delaying the decision - You want to be a software engineer and think a PhD will open more doors (it often doesn't, and can actually signal overqualification for many roles) - You're doing it for prestige — the opportunity cost is enormous and the prestige doesn't cash out the way you might think
Industry PhD programs (where companies sponsor your PhD) exist at a handful of companies (Microsoft, Google, etc.) but they're rare and highly competitive. Don't plan around them.
The median PhD in CS takes 5.5 years. You will spend years working on a problem that may not pan out. You'll be paid significantly less than your BSc peers during that time. Your social life will be constrained. If you love research, this is a worthwhile trade. If you don't, it's genuinely not.
UofA's Actual Strengths for Grad School¶
UofA is not MIT. Be clear-eyed about that. But it has genuine world-class strengths that are worth knowing.
Reinforcement Learning — Genuinely World-Class
The Reinforcement Learning and Artificial Intelligence (RLAI) lab is, arguably, the best place in the world to study reinforcement learning. Rich Sutton co-wrote Reinforcement Learning: An Introduction — the textbook used in RL courses everywhere, available free online. Michael Bowling, Martha White, Adam White, Csaba Szepesvári are globally recognized researchers. DeepMind has deep ties to this lab; ideas from RLAI influenced AlphaGo and subsequent work. If you want to do RL research, Edmonton is the right city.
Amii — Well-Resourced AI Ecosystem
Alberta Machine Intelligence Institute (Amii) is UofA's hub in the Pan-Canadian AI Strategy (alongside Vector Institute in Toronto and MILA in Montreal). This means funding, industry connections, and a serious research ecosystem around ML. Grad students here have access to resources that many other Canadian universities don't.
The Hinton Network
Geoffrey Hinton (Turing Award winner, father of deep learning) was a long-time collaborator with UofA researchers. This creates a network of connections to Toronto, Google, and the broader deep learning community. It's not just a name-drop — it means real academic relationships and collaboration opportunities.
Databases — M. Tamer Özsu
Tamer Özsu is a legend in database systems, particularly distributed databases and NoSQL. If you're interested in data systems research, he's one of the best supervisors in Canada.
Other solid areas: networks, operating systems, HCI, NLP (growing). These aren't RLAI-level famous but they're legitimate research groups with opportunities for graduate students.
The Application Process¶
GPA: 3.5+ is where you become competitive. 3.7+ makes you a strong applicant. Below 3.3, you'll need exceptional research experience to compensate.
Research Experience: This is the most important factor after GPA. A thesis-based MSc or PhD requires you to demonstrate you can do research — not just take courses. NSERC USRA or significant volunteer RA experience is what makes your application credible. Letters from professors who supervised your actual research (not just course instructors) are worth far more than letters from professors whose courses you aced.
GRE: UofA CS does not require the GRE. Verify this each year as policies change, but as of recent cycles it has not been required.
Statement of Purpose: This is the most important document in your application after your transcript. Be specific. Name the professors you want to work with and explain why — not just "I'm interested in machine learning" but "I want to work on continual learning because I think the stability-plasticity dilemma hasn't been adequately addressed in the online RL setting, and Dr. White's work on this is the most promising direction I've seen." Generic statements get generic consideration.
Reference Letters: You need three. They should ideally come from professors who supervised your research or who know your academic work in depth. A letter from a professor whose large lecture course you got an A in is less valuable than a letter from a prof who worked with you for a semester in their lab.
Deadline: Typically December–January for a September start. For international programs (if you're considering leaving UofA for grad school), October–December is more common. Check each program's specific deadline.
Funding¶
Funded MSc at UofA: Expect roughly \(20,000–\)28,000/year as a stipend, drawn from your supervisor's research grants. This is livable in Edmonton (lower cost of living than Toronto or Vancouver). It covers rent and basics; you won't be saving much.
Funded PhD: Similar range, sometimes slightly higher depending on supervisor grants and additional funding.
NSERC Postgraduate Scholarships: - PGS-M (Masters): $17,500/year for up to one year. Apply in September of your final undergrad year or first grad year. Highly prestigious — holding an NSERC graduate scholarship signals strongly to future employers and future PhD supervisors. - PGS-D (Doctoral): $21,000/year for up to three years. Apply at a similar time point in your PhD trajectory.
These are competitive national awards. If you receive one, your supervisor no longer has to pay you from their grant, which makes you a much more attractive student to take on.
Alberta Scholarships: The province has additional graduate funding. Check the Alberta Graduate Excellence Scholarship (AGES) and similar awards through UofA's Faculty of Graduate Studies and Research.
One important note: Don't go unfunded. If a supervisor offers you a grad school position with no funding and no plan for funding, be very cautious. Funded positions are the norm in CS at research universities. An unfunded position usually means the supervisor has limited commitment to your success.
The Honest Bottom Line¶
Industry salaries for CS grads have grown faster than grad school stipends over the last decade. A strong BSc grad with two good internships can enter industry at \(90K–\)130K CAD. A funded MSc student makes $20–28K for two years. The financial opportunity cost is real.
That said, money isn't the only variable. If you love research, if you want to work at a research lab, if you're aiming for ML roles at top companies, or if you want to be a professor — grad school is the right path. Be honest with yourself about why you want to do it.
The worst reason to do a Masters: because you're scared to enter the job market and two more years of school feels safe. It delays the problem and adds debt (or at minimum, foregone income) to it.
The best reason to do a Masters: because there's a specific thing you want to learn how to do that requires the research environment to learn it.
If you're not sure, spend a semester as a volunteer RA. If you love it, consider grad school. If it feels like a grind, you have your answer.