Data engineer career path
Data engineering has a long hands-on ladder. Senior, staff and principal data engineers own platform design, data contracts and warehouse cost, so staying technical doesn't mean standing still. Sideways moves go toward the people using the data (machine learning) or down into the infrastructure under it (data platform). Management jobs exist, but data teams tend to be small, so there aren't many.
Updated Pay and job counts from live US job listings
What comes after data engineer?
A common next step for a data engineer is data architect, usually 6-8 years in. You can also step up to data engineering manager, or move sideways to machine learning engineer or data platform engineer. Some change careers and become a technical product manager.
Data engineer jobs in the US
- Advertised pay
- $130K-$136K
- Open jobs
- 17,145
- Median
- $134K
Jobs are counted by title on US job boards. Pay is the middle half of the 21 listings that state a salary.
Where data engineers go next
5 moves people make from this role. The job counts and pay are for each destination role, so you can see what the market wants right now.
| Move to | Type | Usually when | Open US jobs | Advertised pay |
|---|---|---|---|---|
| Data architect | Step up | 6-8 years in | 3,756 | $165K-$178K |
| Machine learning engineerIts career path: machine learning engineer | Sideways move | 3-4 years in | 4,176 | $177K-$216K |
| Data platform engineer | Sideways move | 3-5 years in | 963 | $121K-$148K |
| Data engineering manager | Step up | 5-7 years in | 1,068 | $143K-$182K |
| Technical product manager | Career change | 4-6 years in | 1,760 | $140K-$177K |
Job counts and pay from US listings, September 29, 2026. Pay shows only where at least five listings state it.
What each move takes
Why each move fits, the skills hiring managers look for, and the proof to have on your resume before you apply.
Data architect
Step up3,756 open US jobs$165K-$178K advertised
Anyone who has lived through a messy migration knows which storage and governance decisions are expensive to undo. Data architects make those decisions up front, across every part of the business.
- Usually when
- 6-8 years in, usually after leading a warehouse or lakehouse migration
- Skills to add
- Open table formats (Apache Iceberg or Delta Lake)
- Data catalogs and governance (Unity Catalog or DataHub)
- Data contracts between producers and consumers
- Databricks Certified Data Engineer Professional or Google Professional Data Engineer
- Have this on your resume first
- A platform or modeling design you led across several business areas, with the cost, reliability or adoption outcome.
Machine learning engineer
Sideways move4,176 open US jobs$177K-$216K advertised
Models break most often because of the data, through stale features, bad training sets or leaked future values. You already own that part.
- Usually when
- 3-4 years in, especially if you've built feature or training pipelines
- Skills to add
- scikit-learn and PyTorch fundamentals
- Feature stores such as Feast
- Model training and evaluation
- MLflow
- Have this on your resume first
- Feature or training pipelines you built for a model that runs in production.
Data platform engineer
Sideways move963 open US jobs$121K-$148K advertised
Platform work turns one-off pipeline fixes into infrastructure everybody uses, like scheduling, compute, access control and CI for data. Pick it if the systems interest you more than the business logic.
- Usually when
- 3-5 years in, once you're building shared tooling rather than one pipeline at a time
- Skills to add
- Kubernetes
- Terraform
- Data quality testing with Great Expectations or Soda
- Warehouse cost monitoring and workload management
- Have this on your resume first
- Shared data tooling other engineers adopted, with a reliability, cost or onboarding improvement.
Data engineering manager
Step up1,068 open US jobs$143K-$182K advertised
Data teams serve many internal customers at once, so senior engineers already negotiate priorities and service levels. As a manager, that negotiation and growing the team is the job.
- Usually when
- 5-7 years in, typically after leading on-call and juggling requests from several teams
- Skills to add
- Intake and prioritization across stakeholder teams
- Defining data SLAs and on-call practices
- Hiring and coaching engineers
- Budgeting warehouse and tooling spend
- Have this on your resume first
- A cross-team data project you coordinated, with the engineers involved and what got delivered.
Technical product manager
Career change1,760 open US jobs$140K-$177K advertised
Data platforms need PMs who understand where data comes from, how fresh it is and what a schema change breaks downstream. You've been translating between the teams that produce data and the teams that use it for years.
- Usually when
- 4-6 years in, if you've been the one negotiating requirements with the people who use the data
- Skills to add
- Writing PRDs for internal platform products
- User research with analysts and data scientists
- Adoption and usage metrics for internal tools
- Roadmap prioritization
- Have this on your resume first
- A data product or platform feature you scoped with its users, and how many teams adopted it.
Now check your own resume
Your own resume will give you a sharper answer than this page. Your years, tools and wins change which move fits.
The suggested roles and fit scores come from AI. The job counts and pay come from live listings.
Questions data engineers ask
What comes after senior data engineer?
Usually staff or principal data engineer, data architect, or data engineering manager. The hands-on roles own platform-wide decisions like table formats, contracts and cost. Management owns the team and its priorities. Plenty of engineers move sideways at this point too, into ML or platform engineering.
Can a data engineer move into machine learning?
Yes, and it's one of the smoother moves in data. You already build the pipelines models depend on, so you'd add modeling basics, evaluation and serving. Start by owning the feature pipeline for a model that already exists, so the move has a track record behind it.
Will managed tools make data engineers less necessary?
They've replaced the routine part, mainly writing connectors and scheduling scripts. What's left is the harder work of modeling, contracts, quality, cost and reliability at scale. Engineers who only moved data from A to B feel the squeeze. The ones who own what the data means don't.