I wrote this in 2019, a few years into my data science career. By then I had been through several job searches of my own and had also helped hire other data scientists. Data science was at the height of its hype, and plenty of people wanted in.
These are the ten tips I gave aspiring data scientists at the time. I have kept them as they were, with a little more explanation, and added a note at the end on what has changed since.
On your resume
1. Show how your work added business value
Do not write something like “Automated logistic regression model to speed up loan conversions.”
Write something like “Automated logistic regression model to increase loan conversion by 34% and originations by $12 million.” That tells a recruiter that a simple model you built had a direct, positive effect on the business.
2. Do not just list technical skills
A bare list of skills makes recruiters skeptical. Say how you used each one in your past work.
In the interview
3. Know your statistics from a business point of view
Topics I was asked about in interviews included distributions, minimizing loss functions, generalized linear models, cross-validation, experimental design and model performance metrics.
The best candidates I saw knew how these ideas applied to real business problems. You can tell who has actually wrestled with building models by how they reason about a business question. Expect questions like these:
- Why did you optimize for AUC instead of accuracy?
- Is dropping a set of complex features worth the loss in accuracy if it saves on computation?
- How did your stakeholders respond to the data product you built?
4. Be ready to code, solve a problem or work through a case study
This is where you show you can do what your resume says. These exercises usually test concepts from your studies or check that you can handle simple tasks. They are also a chance for the interviewer to see how you think. Do not be afraid to think out loud.
5. Focus on business value, not jargon
I cannot say this strongly enough. You need to be able to explain how your work created value without getting lost in the technical details.
6. Be ready to explain complex ideas simply
Technical and non-technical audiences both appreciate an explanation that is accurate and easy to follow.
Questions to ask them
7. Ask how the company plans to use data science in the long term
In 2019, many companies were still riding the wave of adding data science to their business. A major red flag was a company with no plan for using machine learning or AI in its business, service or product. This matters most if you want to work on data science projects early.
8. Ask about the team’s experience and who you will report to
Will you work with other data scientists? Will you report to someone experienced in data science? These questions affect your growth and your working relationships.
I have found it very rewarding to report to someone with deep data science knowledge. I have also found it rewarding to be the only subject matter expert. Work out which of those suits you.
About yourself
9. Admit the gaps in your knowledge with confidence
Saying what you do not know, and how you plan to cover it, shows that you do not pretend to know everything. What you need to show is the drive to learn.
10. Know your niche and ask about growth
You do not have to be a data scientist forever. The skills carry over to software engineering, product management, team management and business intelligence.
Whether you are just starting or have years behind you, know what you are good at, so you can focus on improving the rest.
Looking back from 2026
Most of these tips still hold, and a few matter more than they did. Tip 1 is now close to a requirement. Hiring managers see many resumes that list the same tools, and many of those resumes were polished by an LLM. A specific number tied to a specific business result is one of the few things that still stands out, as long as you can explain where it came from.
Tips 3, 5 and 6 have become the core of a good interview. Anyone can get a working model or a clean SQL query from an AI assistant in minutes. What an interviewer wants to know is whether you can tell when that output is wrong, choose the right metric for the problem, and explain the trade-off to someone who does not care about the method.
Tip 4 has changed shape. Take-home assignments are harder to judge when candidates can use AI tools, so more teams now do live exercises, ask you to talk through a solution you already submitted, or let you use AI openly and watch how you check its work. Thinking out loud is still the right advice.
The entry level is where the biggest change has been. Much of the routine work that junior data scientists used to learn on, such as first-draft queries, data cleaning and standard reports, is now partly automated. That has made junior roles fewer and more demanding. If you are starting out, domain knowledge and a real project you can talk about in depth count for more than a longer list of tools.
Tip 7 is now about how a company uses AI, not whether it does. Almost every company has an AI plan on paper. The better question is what data it is built on and who owns the results. Tips 8 to 10 have not changed at all.