data science was dying 7 months ago?
It was also dying 2 years ago.
And dying 3 years ago.
And not to mention it was also dying 5 years ago.
However, from where I stand, this is definitely not the case. People still seem to land data scientist jobs.
I mean, I literally help people do this every week in my coaching programme.
So, what on earth is going on?
Well, in this article, I want to break down:
- What the current data market looks like
- What it actually means to be a data scientist
- And, what you should be doing to land a job in this current climate
Let’s get into it!
Market Outlook
As many of you will know, there were significant layoffs during 2022 and 2023, with nearly 90,000 tech employees being laid off in January 2023 alone.
In fact, it was so severe that TechCrunch even created an archive of all the layoffs that occurred during this period!
However, according to a study by 365datascience, data jobs were not that affected by these layoffs; they found that:
Interestingly, our sample’s largest group of laid-off employees did not hold tech jobs — 27.8% worked in HR & Talent Sourcing, while software engineers came in second with 22.1%. Marketing employees followed them with 7.1%, customer service with 4.6%, PR, communications & strategy with 4.4%, etc.
For example, only 2.7% of people laid off from Amazon during this period had the title of data scientist.
According to another study:
Data science job postings grew 130% year over year after hitting rock bottom in July 2023, while data analyst openings grew 63% in the same time period.
And we can also see that the salary of data jobs as a whole has been growing over the years.
So, it’s clear that data science is not dying whatsoever; if anything, it’s growing.
However, why does it feel very hard to get a data scientist job right now, especially at the entry and junior levels?
To explain that, we need to look past the numbers and really understand what the modern data scientist is.
Data Science Evolution
As an insider in this field, let me tell you a secret.
Data science is not dying; it’s evolving.
10 years ago, companies would hire data scientists to tinker with machine learning models in Jupyter Notebooks.
In fact, this is exactly what my first data science job was like.
A data scientist was like a Swiss Army Knife — one person expected to do everything from cleaning data to building models and presenting to the CEO.
However, over time, companies realised they were getting no return on investment from this strategy, so they became more stringent about roles and responsibilities to ensure they were not wasting their money.
This has led the data science job to become fragmented, and the title has become meaningless, as you will find data scientists doing completely different jobs at different companies.
In general, three flavours of data scientists exist today.
Analyst
This type of data scientist is closely aligned with the business side and primarily focuses on reporting workflows and experimentation.
For example, you would:
- Get data from a company database or other sources.
- Write some code that is very linear and bespoke by nature, starting with ingesting data, cleaning it a bit, then doing some EDA and some inferential or basic modelling work.
- Once complete, you put together a report that details the analysis, provides visualisations and other metrics, and offers a recommendation based on the analysis’s goals.
This type of data scientist is more of a data analyst and typically requires more business domain knowledge.
Engineering
The focus of this type of data scientist is on building and deploying solutions. This can be a range of things like:
- Internal software tooling
- Machine learning models that drive decision making
- Building libraries
This role leans more toward software engineering, but unlike a software engineer, it requires greater knowledge of maths, machine learning, and statistics.
Nowadays, this type of job has moved beyond the “data scientist” title and is now called a machine learning engineer.
This is not entry level position, and normally requires 2–3 years experience in an adjacent role like a software engineer or analyst first. So many graduates and people with little experience would struggle to break into this specific data science position.
Infrastructure
This type of data scientist is the rarest, mainly because it has its own title: data engineer.
The goal of this role is to build the data infrastructure and pipelines to house the business’s data. This data is then used downstream by machine learning engineers, analysts or even non-technical stakeholders.
This role has become increasingly important, especially with the emergence of generative AI in recent years, which requires the ability to effectively store large amounts of data and stream it with low latency.
At some companies, you may also be an analytics engineer, which is a more business-focused data engineer.
I know, so many titles, its hard to keep up!
Junior vs Senior
A study published in September 2025 has been making quite a few waves in the data and machine learning space.
The study examined 285,000 companies between 2015 and 2025 and how their adoption of GenAI has affected their hiring processes for junior and senior positions.
Note: this applies not just to data scientist jobs but to all jobs at these companies.
You can see in the plot below that hiring for senior positions is still increasing, whereas hiring for junior positions is decreasing.
This makes intuitive sense, as juniors’ responsibilities are likely easier to automate with AI than seniors’ due to the wealth of experience they have built over the years.
What I want to make clear, though, is that companies aren’t making juniors redundant nor are there no more junior positions left on the market.
Most people will look at this graph and think that the junior data science market is becoming extinct. But that is objectively not the case.
Hiring is still happening, but the rate of new positions being posted is not increasing. The supply curve remains unchanged while demand remains high.
That’s why it feels so hard to get an entry-level job nowadays.
What Can You Do?
I am going to be honest, it is becoming more competitive to break into data science, but it’s not impossible.
Gone are the days when all you needed was basic Python and SQL, and having done Andrew Ng’s Machine Learning course.
These are things everyone has nowadays, so you need to go the extra mile and differentiate yourself more than you used to.
There are many ways of doing this, for example, you adopt and specialise in certain technical domains like:
- GenAI
- Model deployment
- Time series forecasting
- Recommendation systems
- Domain-specific expertise
Specialists are arguably becoming more important as knowledge is increasingly democratised by AI. Having deep expertise is almost a rarity nowadays.
Another option is to go for a lower-level position, like a business or data analyst role, that is more friendly to junior and entry-level positions, and then slowly build your way up to a full-time data scientist position.
You should also focus on areas that AI can’t really replace:
- Communicating effectively with different audiences
- Understanding the business impact of your work
- Critical thinking and knowing what problem to solve
- Strong fundamentals in maths and statistics
- Relationships and network
These are timeless skills, especially the last one.
You might have heard the saying:
It’s not what you know, but who you know
I actually disagree with this.
The real power is in who knows you.
If you have a solid network and relationship with many people in the field who value and trust you, you can tap into this to get referrals, opportunities, or even expand your network further.
The leverage this provides is incredible. I always tell my coaching clients that referrals and networks are literally the golden ticket to getting top-end data science jobs.
And all it requires, is just effort and pushing yourself out of your comfort zone to speak to people you want to connect with.
Technologies will come and go, but actual human relationships will remain central for your whole career.
The truth is, you are going to need to reinvent yourself every 3–5 years as a data scientist, since technology shifts very quickly.
So asking “Is data science dying?” misses the point.
Data science is always technically dying as it’s consistently evolving and transforming.
But that’s what makes it exciting.
And if you are willing to up-skill and put in more effort than others, you will be rewarded very well.
If you’re ready to dive into data science after reading this, that’s a great first step.
But here’s the reality: I’ve been in this field for five years, and looking back, I spent my entire first year on tasks that were a complete waste of time. In today’s hyper-competitive market, you don’t have the luxury of trial and error.
To avoid my mistakes and speed-run your progress, check out this guide where I map out exactly how I would become a data scientist again.
Another Thing!
Join my free newsletter where I share weekly tips, insights, and advice from my experience as a practising data scientist and machine learning engineer. Plus, as a subscriber, you’ll get my FREE Resume Template!
Dishing The Data
Weekly emails helping you land your first job in data science or machine learningnewsletter.egorhowell.com