The framework I used
Map the topics that actually get tested, put them into one master doc, then do live mock interviews with AI.
Open on LinkedIn →
← Crocodata
Start here
Part 1
Part 2
The framework
Other paths
Free guide · PhD → industry
I finished my PhD with no job offer.
Four years of theoretical physics at the Max Planck Society in Berlin. Then six months at Stanford, with Silicon Valley right outside the door: everyone building something, everything moving fast. I wanted to be part of that momentum. So when I handed in my thesis, I didn't stay in research. I started applying.
Half a year and around 100 applications later, most of them rejections, I started at Google as a data scientist. This page is the framework that got me there: what I did, in what order, and the exact doc I studied from. Take it, copy it, make it yours.
4 yrsPhD at Max Planck, Berlin
~100applications in 6 months
5Google interviews
Sep '25first day at Google
How I landed the Google job
Two short videos where I walk through exactly how I prepared. Watch them first, then use the framework below as your checklist. The videos load from LinkedIn only when you press play.
5 interviews, prepared with AI
Three stats and data intuition rounds, one coding, one Googleyness. How I prepared for each, and why the interviewer is a collaborator.
Open on LinkedIn →More on Instagram: my journey from academia to Google · @what.the.tech.is.going.on
Three phases, six habits
Build things, close the gaps, then treat interviewing as a skill you train. That's it. The details are below.
01
Build proof
During the PhD, if you canYour PhD shows you can think. Side projects show you can build things industry cares about. Start them early, even small ones.
Always have a side project
I built a quant trading bot. You don't have to. Build whatever pulls you in: a small app, a data analysis on something you care about, a tool for your lab.
What matters is that it's real, finished enough to show, and close to the jobs you want.
Give every project a story
Projects open doors, stories get you through them. For each one, know:
- The problem and why you cared
- What you decided and why
- What went wrong and what you learned
- The result, ideally with a number
02
Close the gaps
Before and while applyingRead ten job descriptions for roles you want. Write down every skill you don't have yet. That's your curriculum. The physicist in me treated it like an experiment: find the gaps, test, repeat.
Take the courses
Coding, plus a lot of random courses: whatever kept showing up in the job ads I wanted. No course is too small if it closes a real gap.
Learn by doing
Feed what you learn straight back into your side projects. A course certificate is nice; a project that uses the skill is proof.
03
Train interviewing like a sport
The application phaseInterviewing is a skill, and skills need reps. I didn't start with my dream job. I started where a bad interview wouldn't hurt.
Go broad first
Apply widely at the start, including jobs you're less excited about. I tried quant trading, startup accelerators, consulting and data science. The early ones were my practice ground.
Study for every interview
Every single one, same routine: map the topics that get tested, write them into one master doc, then practise out loud until it feels automatic.
Make AI your interview partner
Feed it everything: the job ad, your CV, your project stories, your master doc. Then let it interview you out loud. It asks follow-ups and finds your gaps before the company does.
You are my interview partner for this role. Here is everything: [job description] [my CV] [my side projects and the story behind each] [what I know about the company and the interview format] 1. Tell me what this interview will most likely test, and where my gaps are. 2. Make me a study plan for the days until the interview. 3. Then interview me: one question at a time, wait for my answer, grade it honestly, show me a stronger answer, and move on.
In the room: treat the interviewer as a collaborative friend. The questions are open-ended and creative, and there is no perfect answer. They want to see how you think and work.
The loop: apply broadly → study for each interview with AI → interview → write down what you got asked → feed it back into the next round. Every interview makes the next one easier.
And protect the basics
It's a painful time, and it can take months. Don't forget to take care of yourself. Here are my non-negotiables →
- Working out, even when exhausted
- Off-screen time
- Time for your passions and family
- 8 hours of sleep, whatever the deadline
Keep at it, and don't sell yourself for less than you're worth.
There are as few excellent positions as there are excellent people, so it's just a matter of finding your match.
Free resource
My Google Data Scientist prep doc
Everything I studied for the Google Data Scientist interviews, in one document: formulas, worked examples, code and answer templates. Here's what's inside.
LogisticsThe interview format round by round, what they score you on, and practical tips.
StatisticsA one-page cheat sheet, probability, distributions, hypothesis tests, resampling.
Modelling & experimentsRegression, A/B testing, causal inference, data quality, survival and time series.
ProductChoosing metrics, product-sense frameworks and interview answer templates.
CodingPython data structures, pandas, SciPy recipes, SQL and worked exercises.
BehaviouralA STAR story bank, Googleyness, a question bank and further reading.
Open the prep doc →
Written for data science, but the logistics, behavioural and coding parts work for most technical interviews.
Insights from friends
My way isn't the only way. I'm collecting insights from friends who made it into tech: what they did, what worked, what they'd skip. More coming soon.
My friend Alex · Data Sovereignty Manager, TikTok
Ex-Activision Blizzard, Microsoft and Tencent. Her three tips for standing out:
- Get a specific referral. Your referrer should describe how you worked together and what you solved.
- The 30-minute hack. Research the parent company and its side projects, not just the main app.
- Day-1 mentality. Show you're curious, ready to solve problems and care about the product.
Coming soon
From a PhD to industry research.
Coming soon
From a PhD to something completely different.
1:1 calls with Alexandra
Want help applying this to your situation?
I've been there twice: academia to Google, and now Google to a startup. Book a call and we'll look at your situation together. You'll leave with an honest outside view and concrete next steps, not a pep talk.
15 minutes
Quick question
$29
One specific question you want a straight answer to: is my CV readable for industry, should I take this offer, does this idea make sense?
Book 15 minutes30 minutesMost popular
Career transition call
$49
Time to go through where you are, where you want to go and what's actually holding you back. We leave with a plan for your next move.
Book 30 minutesWhat we can talk about
- Leaving academia (PhD or postdoc) for industry, and how to explain your research so companies get it
- Picking side projects and turning them into stories that land in interviews
- Getting into big tech: what the interviews and the day-to-day actually look like
- Moving from a big company to a startup, and whether it's the right call for you
- Feeling stuck: you know something has to change, but not what or how
How it works
Pick a time that works for you and tell me briefly what's on your mind.
Pay securely by card. The booking is confirmed right away.
You get a calendar invite with a video link. We talk.
Who you'll talk to
Alexandra Maxi Dudzinski, PhD. Physicist turned tech person: academia, then Google, now a startup. I also run Crocodata, a small studio where I build and publish my own apps, and I make short videos about what's going on in tech. LinkedIn →
Coming next: courses and more videos that go deeper into each step of this framework.
Calls are informal mentoring conversations, not therapy, legal or financial advice. Booking and payment are handled by Cal.com and Stripe; the booking window loads from Cal.com only when you click a booking button. Questions? contact@crocodata.net