After Criteo: Target Better

What I took from Criteo comes down to two concrete things.

The first: my Python level. Before Criteo, I had five days to learn the language — five days of full algo, no time to actually understand Python. Too short, too scattered.

The second: the format didn't suit me. Full algo, no camera, no interaction — no way to show how I think or how I approach a problem.

The lesson: learn Python properly, and find companies whose process lets me show what I have to offer.

Why Alan

Alan is a French health tech company whose mission is to make healthcare accessible. That's what drew me in first — a company with a concrete mission, in a sector that matters, with real growth.

Digging further, I found their recruitment process. Alan publishes everything — real articles on their engineering blog, Medium posts written by Alaners, a dedicated page on their site. Not an HR summary. The stages, the spirit behind each one, what each interview is broadly about.

Reading that, I understood the technical interview wasn't pure algo. Business logic, Python, CoderPad. Explaining your approach, structuring your code, writing tests — explicitly what they're looking for.

I didn't know in advance whether it would match. But the signals were clear enough to make it worth genuinely trying.

10 Days to Learn Python

This time, no scattering. Ten days, one goal: Python oriented toward business logic.

I built my own exercises based on what I'd understood of the Alan format — that's the sandboxPython repo on my GitHub. Log parsing, API calls, data manipulation, SQLite. Some OOP to structure code cleanly. pytest for tests, because Alan states it explicitly. CoderPad on the side to get used to the live format — coding while explaining what I'm doing, structuring my approach out loud, iterating rather than chasing a perfect solution from the start.

The Process

Once I submitted my application, I heard back quickly. What struck me immediately: after each stage cleared, Alan sends an email with concrete information about the next one. No ambiguity, no waiting in the dark. You know what to expect.

Screening — Stage 1

A first conversation with a recruiter, 45 minutes. Not technical — a discussion to check motivation and introduce the role.

How it felt: stressed, but it went well. I was prepared — I knew Alan's mission, I knew why I was applying, I'd read their documentation in detail. When you've done the work, it shows.

Technical Interview — Stage 2

CoderPad, business logic in Python, about an hour. What it really tests: the engineering mindset. What's evaluated isn't just whether the code runs — it's how you frame the problem before writing anything, how you communicate during, whether you write tests, whether you know how to iterate. Alan wants to see a process, not a magic solution.

How it felt: intense. Live coding stays stressful no matter how prepared you are. But this time I had the right instincts — clarify the problem before coding, explain my choices, write my tests. The format gave me room to show how I work.

Cultural Fit — Stage 3

This interview is different from the first two. Centered on you, your way of being and working. Would you be someone they could work with day to day? No trick questions. Your motivation, your past experiences, how you react in certain situations, what you're really looking for.

How it felt: the most uncomfortable stage, because you can't prepare for it the same way.

You can prepare code. You can't prepare who you are.

But it was also the stage where I felt most at ease — the questions made sense, and I had real answers to give.

Team Interview — Stage 4

A conversation to see if it works, on both sides — human and professional. You talk about concrete things, how the team operates, what it actually means to work there.

How it felt: the stage that surprised me most, in a good way. Honest and direct.

That same evening — I got in.

What I Took From It

My process took six weeks. That's long — but I needed that preparation time. Being methodical: understanding what was evaluated at each stage before going through it, building my preparation accordingly, not improvising.

That doesn't mean planning everything. The fit, you can't prepare for. But the rest — Python, the live format, knowing the company — that's work, and work pays off.

Criteo didn't work out. I understood why. I fixed what could be fixed and applied elsewhere.

Six weeks later, I had an internship at Alan.