↵ select ↓ ↑ navigate esc close

Do Data Roles Have a Future?

Joe Reis ·

Does data engineering have a future? I get asked this all the time now, along with similar questions about pretty much every other data role. While things are still playing out, here are my thoughts on where we stand today.

Students wonder if they should bother getting into data engineering, experienced folks worry if their jobs will be around in five years, and vendors keep pitching AI agents that promise to replace entire data teams. Having co-written Fundamentals of Data Engineering, I’ve been thinking about this a lot.

The short answer is yes, data engineering has a future. The data engineering lifecycle and its undercurrents aren’t going away. What’s far less clear, though, is whether the tactical day-to-day work we’ve traditionally associated with data engineers will still define the job. What happens when the tasks we spent years mastering suddenly become dirt cheap to automate? As a result of these shifting economics, people are facing a major identity crisis, from analysts and analytics engineers to data scientists, architects, and managers alike.

The Shifting Economics of Technical Artifacts

The economics of data work are shifting fast. Writing SQL, dbt, or Python, and managing infra are increasingly being handed off to AI agents that can generate code, run tests, write docs, and debug systems faster and cheaper than humans. I don’t know anyone strictly hand-writing code these days. And that’s not necessarily bad, since AI speeds things up and lets us experiment rapidly with implementations that used to take days. Rapid iteration is the name of the game nowadays.

What’s really changing is how we generate technical artifacts, including code, pipelines, dashboards, and docs. But artifacts alone don’t solve business problems. To figure out where value actually lies, it helps to break data work into three parts:

  • Production. Building the artifacts - writing code, creating pipelines, building dashboards, and writing documentation.
  • Interpretation. Understanding the business context - figuring out what data actually means, making architectural choices, and uncovering why metrics don’t line up.
  • Accountability. Deciding what’s right, taking on risk, and owning the outcomes when things break.

AI makes production dramatically cheaper, but navigating ambiguity and owning big decisions remain fundamentally human and organizational challenges. This is the messy part that AI alone won’t solve. But I’m hopeful AI will shine a bright light, expose problems in real time, and help humans resolve them much faster.

Even as models get better at handling complex tasks, the immediate takeaway is clear: the balance between production, interpretation, and accountability has changed for good. When pumping out artifacts becomes almost free, knowing which ones to build, how they ought to behave, and how to verify their accuracy become the real core skills.

Findings from the 2026 Practical Data Community Survey showed that while 80% of respondents used AI daily, only about 10% had actually built it into organizational workflows, and nearly 89% were still wrestling with basic data modeling problems. That gulf speaks volumes. Code generation tools are moving at warp speed, but organizations still struggle with semantic clarity and data modeling. Much to people’s surprise (and vendor’s dismay), things like data modeling and ontologies are sociotechnical problems that can’t simply be automated away.

Take revenue calculations, the classic trope we like to debate when modeling data. AI can easily spit out syntactically correct code, but it can’t automagically determine whether you mean gross vs. net or recognized vs. invoiced revenue. Without solid data modeling and domain clarity, automated tools will just generate incorrect business logic faster. This isn’t bad in itself, since finding things out faster can help resolve issues more quickly than in the past. But we still need to resolve the ambiguity, and that’s a human problem.

AI can automate writing code, but it can’t settle disagreements between stakeholders over definitions. Human judgment is still required to set policy and ensure systems align with what the business actually needs. This is why decades of IT implementations still fail for the same reason: sociotechnical work is very messy and difficult.

Roles are Blurring. Titles Won’t Matter as Much

Every data role is shifting in response to these changing economics. Here’s a breakdown of where I see roles heading.

  • Data engineers will spend less time plumbing pipelines by hand and more time on system design, reliability, cost management, and orchestrating automated workflows.
  • Analytics engineers will spend less time churning out dbt models and SQL transformations and more time refining business semantics, validation, and data contracts.
  • Data analysts will move away from static reporting toward asking sharper strategic questions, digging into edge cases, and interpreting subtle business trends.
  • Data scientists will delegate standard modeling tasks to AI, placing more emphasis on experimental design, causal inference, and real-world evaluation.
  • Data architects will design across complex environments, knowledge graphs, and agentic workflows, taking a broader, macro view of data movement.
  • Managers will need to stay hands-on as smaller teams leveraging AI pull off work that used to require big departments, blending leadership with direct technical oversight.

Again, AI is eroding production work. With AI assistance, engineers can easily run complex analyses, analysts can build working tools, and data scientists can push code straight to production. In the bigger picture, I see these roles blurring together. Last year, I noted this convergence. Today, that convergence is accelerating. This convergence is giving rise to versatile generalists who move fluidly across domains. As tooling friction disappears, rigid job titles will matter much less than what you can actually execute. Deep specialists in fields like distributed systems, performance tuning, and security aren’t going anywhere, but for general data roles, performance and execution will trump traditional titles every time. My podcast about this caused some angst in my Discord over the weekend, as people struggled with what future titles would look like, and how to write job descriptions where roles are converging to someone who’s simply a “data practitioner.”

One real concern is how junior engineers will grow. Historically, we built intuition by writing code from scratch, breaking things, and fixing them. As entry-level coding tasks get automated away, companies need to get deliberate about creating pathways for early-career engineers to develop real diagnostic judgment. True engineering seniority comes from understanding how systems break, spotting subtle flaws in logic, and being accountable for production systems. These are skills you simply can’t acquire by blindly accepting AI suggestions.

If you’re starting out or adapting your career right now, focus on mastering core engineering fundamentals while using AI tools aggressively. Read the generated code carefully, understand what’s under the hood, build real projects, and learn to justify your design decisions.

What’s Left For Humans?

Looking ahead, a few main trends stand out.

  1. Leaner, highly effective teams will punch far above their weight class. Big teams simply won’t be necessary and will probably be a liability. This sucks for managers who like to build kingdoms, but it is what it is.
  2. Fluidity across roles will grow as lower implementation costs break down traditional silos. Think about work charts versus org charts, and how companies are delayering their org charts today. This will accelerate as companies get leaner and automate more with AI.
  3. Autonomous agents working as both data producers and consumers will demand stronger governance, data lineage, and semantic consistency. This is where humans can help resolve ambiguity
  4. Validating and evaluating generated code will be just as important as writing it in the first place. As I wrote in the Post Literate Engineer, this will be interesting since humans are reading and writing less while AI reads and writes more.

Even as technical roles adapt, people who can connect business needs with technical solutions will always be in high demand. As routine tasks get automated, we can spend more energy on strategy and solving higher-level problems.

What’s left for humans in data work? So much of our identity has been attached to the production stuff. Countless user groups are dedicated to programming languages and frameworks. The key question for our industry is this: what value do data professionals bring when producing code is no longer the bottleneck? Thriving in this new era means doubling down on fundamentals, rethinking our skill sets, and building strong networks.


In this Freestyle Friday episode, I chat about the various craziness and trends in data and AI right now.

Freestyle Fridays and my other podcasts are available on Spotify, Apple Podcasts, and wherever you get your podcasts. Please support the show with a review. It means a lot.


Mixed Model Arts

Mixed Model Arts is out. If you like the book, please leave a review on Amazon!

Buy it on Amazon ($29.99 Kindle, $49.99 paperback).

Buy the ebook directly ($29.99).

Or subscribe to Practical Data Modeling and get the ebook plus upcoming courses, community events, etc. ($57 for now; the price is increasing soon).


Where I’m At

Here’s what I’ll be doing. Most of the time, I’m giving a talk. Sometimes, I’m just hanging out.

More to be announced very soon.

Find My Other Content Here

📺 YouTube - Interviews, tutorials, product reviews, rants, and more.

🎙️ Podcasts - Listen on Spotify or wherever you get your podcasts

📝 Practical Data Modeling - This is where I’m writing my upcoming book, Mixed Model Arts, mostly in public. Free and paid content.

The Practical Data Community

The Practical Data Community is a place for candid, vendor-free conversations about all things tech, data, and AI. We host regular events, including book clubs, lunch-and-learns, Data Therapy, and more.

🤖 Join on Discord

Thanks for reading! Subscribe for free to receive new posts and support my work.