Starting a PhD

🚧 This section is a work in progress

The most critical aspect of a PhD is pursuing a topic you genuinely enjoy. If you believe you have found a team that does exactly what you are looking for, go for it. A PhD should be a conscious choice driven by passion, not a default path taken because a professor offered you a spot and it felt like the “easy” option. Many students fall into the trap of starting a doctorate out of inertia rather than strategic career planning. If you are not sure about what you like, try to get some experiences ahead of time to work in research (either at your university, foreign university or with industrial research internships). Since a PhD is a commitment that lasts for years, having clear ideas about what you like is highly important: do you want to do AI? If so LLMs? Computer Vision? Fair AI? Or do you want to go for cybersecurity? Another strategy, if you don’t have the chance to get experience earlier, is to read papers. Read influencial and recent papers about the area that you might be interested in and see if the sparkle ignites within you. Once you think you might have an idea of a topic that might interest you, move on to the next part of this guide.

When choosing a PhD path, there are three main categories:

• Academic PhD
• Industrial PhD
• Special Programs

This is the traditional route. While your specific research topic is vital, the environment and the Principal Investigator (PI) you choose will define your future career. Key factors include industry connections, research impact, and the university’s ecosystem.

To illustrate, let’s compare a generic university in a small city (City X) versus a top-tier hub like ETH Zurich.

• Networking: It is highly likely your PI has direct ties to industry giants (Microsoft, Google, Meta, etc.).
• Compensation: Salaries are significantly higher (e.g., 60k–80k+ CHF/EUR).
  Note: While the cost of living is high, the saving potential often remains higher than in low-cost cities.
• Environment: You are immersed in an ecosystem of innovation, surrounded by peers launching startups or joining Big Tech.
• Resources: High funding levels allow for high-risk, high-reward research.

• Compensation: Salaries are generally lower (e.g., ~15k–30k EUR depending on the country).
• Networking: Connections may be limited to local SMEs rather than global leaders.
• Resources: Funding for hardware (GPUs) or conferences may be tighter.

The name of the university matters, but the Supervisor matters more. A famous professor might have 40 students and never speak to you.

The “Toxic PI” Check:
Before accepting, email current PhD students in the lab. Ask:
“How is the work-life balance? Is the PI supportive?”
If they hesitate to answer, it is a red flag.

Unlike the US, where you often do 1–2 years of coursework, in these European programs you usually start research on Day 1.

• ETH Zurich: Applications are generally open year-round (varies by lab).
• EPFL: Two strict rounds per year (deadlines in mid-December and mid-April).
• TUM (Technical University of Munich): Strong industrial ties.
• TU Delft: Excellent for engineering.
• Tübingen / Max Planck Institutes: World-class research centers.
• Imperial College / Oxford / Cambridge: Top tier, though funding structures differ for non-UK citizens.

Strategy: These environments are hyper-competitive. A smart strategy is to apply for a Master’s Thesis position in your target lab first. This allows the professor to vet you before committing to a full PhD offer.

Remember when we talked about the PhD topic before? Now is the time to use it. The number of professors at these univeristies might be huge and they might not work with what you like. Asking them to take you on a topic that they don’t deal with means you didn’t even do the slightiest effort to understand their research interests. Most good universities will have a proper website describing the professors in each faculty and their research interests. Try to reach out to those tailoring the email to their profile, adding why you would like to work with them (maybe read some of their papers) and what experience you have in the field. Do not expect an answer: they are very busy and, if your topic is mainstream, is likely they will receive many emails similar to yours every week. It’s a numbers game, but remember to put in the effort if you want to have some chances.

An Industrial PhD involves conducting research primarily within a company (usually 3–4 years) while affiliated with a university. The work is typically split 50/50 between academic requirements (papers) and company goals.

• Availability: Common in Germany, Sweden, The Netherlands, and France (look for CIFRE contracts in France, which are the standard).
• The “Golden Handcuffs”: Be aware of Intellectual Property (IP). The company usually owns your results, which may prevent you from open-sourcing code or publishing freely. This can sometimes hinder a transition back to pure academia later.
• FAANG: Big Tech companies (Meta, Google) rarely do formal “Industrial PhDs” (they prefer Internships or Residencies). However, many other major firms do.
• Active Companies: IBM, Microsoft, SAP, Huawei, Bosch, BMW, Mercedes Benz, Siemens, ABB, Roche, Novartis.

Europe offers unique networked programs that differ from standard university tracks.

This is a premier opportunity for AI/ML students.

• Mechanism: You apply centrally. If selected, you are matched with high-profile advisors across Europe.
• Tracks: Industrial Track (requires a 6-month internship and industrial mentor) or the Academic Track.
• Benefit: You gain affiliation with the best supervisors in Europe, and the ELLIS brand is a strong signal to recruiters.

I believe the Industrial Track to be very important. If you manage to get an industry supervisor that is in Google, Meta or other important tech companies, then you will HAVE TO spend at least 6 months there, as a requirement of the PhD program itself.

One of the most prestigious funding schemes in Europe.

• Benefit: Extremely high salaries (often well above local standards) and generous budgets for training/travel.
• Requirement: You typically must move to a new country to be eligible (mobility rule).

These are pure research institutes in Germany. They offer immense funding and no teaching obligations, allowing for 100% focus on research.

A PhD is not a prison; it is a platform. You should utilize your time to explore:

• Research Internships: Essential for getting industry exposure and boosting your CV.
• Visiting Positions: Spending 3–6 months at another university to collaborate.

Advice: Do not wait for your supervisor to suggest this. Actively seek these opportunities.

• CS PhD Statements of Purpose is a list with examples of statements of purpose for Computer Science PhD programs. It is a great resource to get inspiration for writing your our proposal. Usually the proposal is a key part of the application process for a PhD program regardless of the country in which you are applying.