Job Description
1. Algorithmic & critical thinking "Walk me through a time you had to choose between two modelling approaches — what was the problem, how did you reason through the trade-offs, and would you make the same call today?"
2. LLM / NLP applied knowledge "Tell me about a project where you worked with embeddings or a language model — what specifically did you build, what didn't work at first, and how did you fix it?"
3. Production-level Python "Describe a piece of code you wrote that someone else had to maintain or build on — how did you structure it and what feedback did you get?"
4. Statistical rigor "Give me an example where your model looked good on paper but you had doubts about it — what metric or assumption made you suspicious and what did you do?"
5. Self-starter mindset "Tell me about a project where nobody gave you a clear brief — how did you define the problem and decide where to start?"
6. Ground-level hands-on work "What's the most unglamorous data or pipeline problem you've had to solve yourself — not delegate, not design, actually fix?"
7. FastAPI / embedding serving "Have you ever exposed a model or embedding pipeline via an API? Walk me through how you set it up and what broke in production."
8. MLOps / CI-CD "How do you currently track experiments and make sure a model you trained last month is still the right one to use today?"
9. Cloud (AWS / GCP / Azure) "Have you used any cloud services for ML workloads — if yes, which ones and for what? If not, how have you been running and storing your work?"
Qualification with Manager
Important things to qualify for Norstella roles please also find questions below.
——————
1. Algorithmic & critical thinking — Must-have Can structure a problem, justify modelling choices, reason independently. The #1 differentiator in interviews.
2. LLM / NLP applied knowledge — Must-have Transformers, embeddings, fine-tuning, RAG — deep hands-on understanding, not surface-level familiarity.
3. Production-level Python — Must-have Clean, testable, readable code. Not notebook dumps. Others can maintain what you write.
4. Statistical rigor — Must-have Hypothesis testing, metric selection, model validation — knows when a model is actually valid vs. just performing.
5. Self-starter mindset — Important Asks questions, challenges assumptions, goes beyond task execution. Not waiting to be told what to do.
6. Ground-level hands-on work — Important 2–5 years applied experience. Not academic, not managerial. Comfortable doing the actual work.
7. FastAPI / embedding serving — Nice to have Bonus points. Experience exposing models or serving embeddings at scale.
8. MLOps / CI-CD — Nice to have Experiment tracking, model registries, deployment pipelines. Helpful, not a filter.
9. Cloud (AWS / GCP / Azure) — Nice to have Cloud-agnostic environment. Any exposure is a plus, no hard requirement.
Project Overview
The client is currently building a new Data Science team (approx. 6–7 people)
The team will be distributed globally (Americas, Europe, Asia)
Our focus is on Europe-based candidates
Location & Setup
Fully remote positions
Open to candidates across Europe (e.g. Poland, Italy, Belgium, UK, etc.)
Cost-effective locations are strongly preferred
Hiring Manager
