Data Engineer Job Description Template 2026

Last updated: June 17, 2026 | By Mike Carter

A data engineer job description should name the specific stack, the seniority level, and a salary band, because the title covers three different roles that pay anywhere from $95,000 to over $175,000 depending on the work. The template below is built from postings that actually filled, not the kitchen-sink JDs that pull 200 applicants and zero hires.

Here is the posting I see most often. “Looking for a data engineer with Spark, Kafka, Airflow, dbt, Snowflake, Redshift, BigQuery, Flink, Terraform, and AWS, Azure, or GCP. Bonus: machine learning experience.” That is not one person. That is a four-person platform team, and the one candidate who could plausibly do all of it is already a principal engineer making $230,000 who is not reading your job board.

So the resumes that come in are a mess. Some are analysts who once wrote a dbt model. Some are backend engineers who touched a Kafka topic. The data engineer you actually need, the one who quietly fits your stack and your stage and would have closed in three weeks, reads the first four bullets, decides the role is either above them or beneath them, and never applies at all. The posting described someone else.

I’m Mike Carter. I run data and engineering searches at KORE1, and our data engineer staffing desk sits inside our broader IT staffing services work. Worth saying upfront, we staff data engineers for a living, so I have a dog in this fight. The template and the salary data below work whether you hand the search to us or run it yourself on a Sunday night with LinkedIn open.

Get the JD right and the search gets shorter. Get it wrong and you burn a month.

Which Data Engineer Are You Actually Hiring?

Data engineering splits into three hiring profiles in 2026: pipeline engineers who build batch ELT into the warehouse, platform engineers who run the orchestration and cloud infrastructure underneath it, and streaming engineers who move data in real time. Each pays differently. Each needs a different posting.

This is not pedantry. It is a $70,000 spread, and the wrong title at the top of your JD sends the search sideways before anyone reads the bullets, because the title is the single heaviest filter on a job board and most candidates decide in or out on that one word alone.

Pipeline / ELT data engineer. The most common hire by a wide margin. This person lives in SQL and dbt. They model raw tables into clean, tested datasets, schedule the loads, and field the Slack message when the finance dashboard looks wrong. Warehouse-centric. Snowflake, BigQuery, or Databricks, with dbt for transformation and Airflow or a managed scheduler on top. They write code, but the code is mostly SQL and Python glue, not distributed-systems work. Entry sits around $80,000 to $100,000. Mid-level lands $105,000 to $130,000. If your problem is “our data is in the warehouse but nobody trusts the numbers,” this is your hire.

Then there is the person who keeps all of that running when it grows.

Platform / infrastructure data engineer. Owns the plumbing, not the models. Orchestration at scale, cloud cost, infrastructure as code, the access controls, the reason a job that took six minutes last quarter now takes forty. Terraform, Kubernetes, the guts of AWS, Azure, or GCP, and a real understanding of why the warehouse bill tripled. This role overlaps with platform and DevOps work, and the comp follows. Mid-level runs $120,000 to $150,000, senior $160,000 to $200,000 in major metros. You need this seat once your pipelines outgrow one person’s ability to babysit them.

Streaming / real-time data engineer. The specialist. Kafka, Flink, Spark Structured Streaming, and the unglamorous reality that “real time” means getting paged when a consumer lags at 2 a.m. Fraud scoring, live inventory, anything where a five-minute-old number is a wrong number. Smaller candidate pool, higher floor. Mid-level runs $130,000 to $160,000, and seniors clear $170,000 and keep going at companies where latency is the product. Most teams do not need this person yet. The ones that do usually know it.

Figure out which of the three you are hiring before you write a word. If a fifteen-minute conversation with the hiring manager does not make it obvious, the posting will describe all three and screen out everyone.

Blended bands from ZipRecruiter (June 2026) and Glassdoor (2026). Coastal metros add 15 to 25 percent. Equity at venture-backed companies can move the senior numbers a lot. Full breakdown in our senior data engineer salary guide.

The Data Engineer Job Description Template

Copy this. Edit the bracketed parts. Delete the role types you are not hiring. I pulled the bones of this from three data engineer reqs we filled this spring, two pipeline roles and one streaming role, then stripped out the parts that were specific to those teams so you can drop in your own. The notes in parentheses explain why each section is there.

Job Title

[Data Engineer / Senior Data Engineer / Streaming Data Engineer / Analytics Engineer]

(Pick the title that matches the work 80 percent of the time. “Data Engineer” is the broadest search term and the safest default. Add “Senior” only if you genuinely require five-plus years, because tacking it on to save face on comp just shrinks your pool. If the role is mostly dbt and warehouse modeling, “Analytics Engineer” is the honest title and a different candidate reads it.)

About the Role

(Two or three sentences. What does this person own, what data, and who do they report to? Skip the company mission paragraph. Data engineers scroll past it.)

[Company] is hiring a [title] to own [the analytics warehouse / our streaming pipelines / the data platform] that powers [what it feeds: product analytics, ML features, finance reporting]. You will build and operate [batch ELT into Snowflake / real-time pipelines on Kafka / the orchestration layer], reporting to [the Head of Data / Engineering Manager / Lead Data Engineer]. This role is [remote / hybrid in {city} / onsite in {city}].

What You’ll Own

(Six to eight concrete responsibilities. Name the actual tools and the actual data. “Ensure data quality” tells a candidate nothing. “Own the dbt test suite that gates the finance models” tells them exactly what Tuesday looks like.)

• Build and maintain [batch / streaming] pipelines that land [source systems: Salesforce, Stripe, app event data] into [Snowflake / BigQuery / Databricks]
• Model raw data into tested, documented datasets using [dbt / SQL], and own the tests that catch a broken model before the dashboard does
• Schedule and monitor workflows in [Airflow / Dagster / Prefect], including the on-call rotation when a load fails overnight
• Partner with [analysts / data scientists / product] to turn vague requests into stable data contracts, not one-off queries that rot in a month
• Manage [warehouse cost / cluster sizing / partitioning] so the [Snowflake / Databricks] bill tracks usage instead of surprising finance every quarter
• Write infrastructure as code in [Terraform] for [the pipeline environments / cloud resources the data platform depends on]
• Own data quality and observability: freshness checks, schema-change alerts, and a real answer when someone asks why a number moved

What You Bring

(Be ruthless about the line between must-have and nice-to-have, because every requirement you add quietly narrows the pool, and most postings pad the must-have column with things that are really preferences nobody would reject a strong candidate over. Make each required line earn its spot.)

• [3-5] years building production data pipelines, with [strong SQL and Python] as the non-negotiable core
• Hands-on experience with [the warehouse you actually run: Snowflake, BigQuery, or Databricks] and a transformation tool like dbt
• Comfort with an orchestrator (Airflow, Dagster, or Prefect) and version control as a daily habit, not an afterthought
• Track record of owning a pipeline end to end, including the part where it breaks and you are the one who fixes it

Nice to have, not required: streaming experience (Kafka, Flink), a specific cloud certification, exposure to [your industry’s data], or experience at [your company stage]. Put these here. Do not let them creep up into the required list, where they quietly cut your applicant count in half.

Compensation and Logistics

(Post the band. I will defend that position in the FAQ. Also state location, work model, and whether visa sponsorship is available, because leaving it out wastes everyone’s first phone screen.)

Salary range: [$X to $Y, based on the bands above and your metro]. Location: [remote / city]. Work model: [remote / hybrid, N days]. Sponsorship: [available / not available].

Tools Worth Naming, and the Ones to Stop Listing

Name the stack you run. Not the stack you read about on Hacker News. A data engineer can tell the difference between a JD written by the team they would join and one written by a hiring committee copying a trend piece, and the good ones quietly close the tab on the second kind.

The transformation layer is settled. dbt is table stakes now, with 44 percent adoption and the title of most-paired tool with Airflow for the second year running, according to Astronomer’s State of Airflow 2026 report, which surveyed more than 5,800 data professionals across 122 countries. If your transformation runs on dbt, say so. It signals that your stack is modern and that the candidate’s existing skills transfer on day one.

On the warehouse, Snowflake and Databricks remain neck and neck in that same survey, with BigQuery close behind. Name yours. “Experience with cloud data warehouses” is filler. “We run Snowflake with dbt, orchestrated in Dagster” is a sentence a strong candidate can picture themselves inside.

One more shift worth knowing. The industry moved from ETL to ELT years ago, raw data loaded first and transformed inside the warehouse, and a JD still built around heavy pre-load transformation reads as dated. It is a small tell. Candidates notice small tells.

Now the part most postings get backward. Listing twelve tools does not make the role look impressive. It makes it look unscoped. When I see Spark and Kafka and Flink and Airflow and dbt and Snowflake and Redshift all marked “required,” I read a team that has not decided what it wants, and so does every candidate worth hiring. Name the three or four tools the person will touch in their first month. List one or two acceptable alternatives. Stop there.

Data Engineer Salary Benchmarks for 2026

The average U.S. data engineer base salary sits around $130,000 in 2026, with most roles falling between $105,000 and $172,000 depending on seniority, metro, and stack. Senior roles routinely clear $175,000.

The aggregators agree on the middle and argue about the edges. That argument is the useful part.

Notice the senior gap. ZipRecruiter pegs the general data engineer near $130,000 and its “senior” sample actually lands lower, around $126,000, because it leans on posted bands that lowball experienced talent. Glassdoor’s self-reported senior number is $175,000. That is a $49,000 disagreement on the exact same title, and it exists because one source scrapes what employers advertise in job ads while the other collects what working engineers say they actually take home each year. When you set your band, trust the paycheck data for senior roles and the posted data for entry.

A caveat on the federal number. The Bureau of Labor Statistics does not track “data engineer” as its own occupation. The closest fit is Database Administrators and Architects, where architects post a $135,980 median, and the adjacent data scientist category is growing 34 percent through 2034. Real data engineering demand sits somewhere in that blend, which is why aggregator data tied to the actual title is more useful for budgeting than any single government code.

From our side of the desk, KORE1 fills data and engineering roles in 17 days on average across 30-plus U.S. metros, and contract data engineers tend to run a premium of roughly 10 to 15 percent over the salaried equivalent, which is worth weighing when you are deciding between contract or contract-to-hire and a direct hire. Our salary benchmark tool will calibrate a range for your exact market and level. If you want the full cost picture, including the parts that are not base salary, we broke it down in our cost to hire a data engineer guide.

Five Mistakes That Kill Data Engineer Job Descriptions

These are the patterns that turn a fillable role into a six-week slog. I see all five most weeks.

1. The kitchen-sink tool list. Already covered, but it earns the top spot because it does the most damage. Twelve required tools is a team’s worth of skills crammed into one req. Cut it to four.

2. Requiring a computer science degree. Some of the best data engineers I have placed came out of analytics teams, physics programs, or a coding bootcamp followed by three hard years of real pipeline work, and not one of them would have survived a strict computer-science-degree filter. Make the degree “preferred” or drop it. Requiring it lops 30 to 40 percent off your pool and buys you nothing a skills screen would not catch.

3. Confusing a data engineer with a data scientist. They are not interchangeable, and a JD that asks one person to build pipelines and train models is asking for a unicorn who will leave in a year. If you are unsure which you need, our breakdown of the data engineer versus analytics engineer roles sorts the adjacent overlap, and the data scientist hiring guide covers the other side.

4. No salary band. In a market this competitive, an unposted range is a self-inflicted wound. More on the why below. The short version is that it costs you good applicants and wastes the screens you do get.

5. “Big data” with no numbers. Saying you work with “big data” tells a candidate nothing. Terabytes a day or gigabytes a month? Batch or streaming? Ten data sources or two hundred? The scale is the job. A data engineer who has tuned Spark jobs over multi-terabyte daily loads is a completely different hire from one who has only orchestrated a few gigabytes of nightly batch, and the numbers in your posting are the only way either of them can tell which job you are describing. State the scale, and the right people will recognize their own work in it.

Questions Hiring Managers Ask Us About Data Engineer JDs

How long should a data engineer job description be?

Aim for 400 to 600 words of actual content. Long enough to scope the role and name the stack, short enough that a candidate reads it on their phone between meetings. The postings that overrun a thousand words are usually padding the requirements list, which is the exact thing that scares off strong applicants. Lead with the role and the stack, keep the responsibilities concrete, and move the legal boilerplate to the bottom where it belongs.

Should the posting require a specific cloud, AWS, Azure, or GCP?

Name the one you run, but treat the others as transferable. A data engineer fluent in BigQuery on GCP will be productive on Snowflake or Redshift within a sprint, because the hard skills are SQL, modeling, and pipeline design, not the specific console. Requiring “5+ years of AWS” when your real need is “knows how to think about cloud data infrastructure” screens out excellent people for a logo. State your cloud, note that comparable experience is welcome, and let the interview test depth.

Do data engineers need a computer science degree?

No, and requiring one is one of the most common ways a good search goes slow. What the role demands is production SQL, solid Python, and a track record of pipelines that survived contact with real data. Plenty of strong data engineers arrived through analytics, a quantitative field, or self-teaching plus a few hard years on call. List the degree as preferred if it helps you sleep. Make the skills the requirement, and your pool gets meaningfully wider without getting worse.

What is the one line that quietly scares off good candidates?

“Must be expert in” followed by a list of ten tools. It reads as a team that does not know what it wants, and the engineers with options move on. Experienced data engineers are pattern-matching for a posting written by people who understand the work. A tight, specific stack signals that. A grab-bag of every buzzword from the last three years signals the opposite, and your best prospects never hit apply.

Contract or direct hire for a data engineer?

It comes down to how settled the work is. A defined project with an end date, a migration or a new pipeline build, fits contract staffing, often at a 10 to 15 percent rate premium over a salaried hire. Ongoing ownership of a platform that the business runs on points to direct hire. Not sure which? Contract-to-hire lets you watch someone work for a few months before committing, which is the honest answer when the scope is still forming. We fill data engineer roles across all three at KORE1, and the right structure follows your roadmap, not the cheaper-sounding option.

How is a data engineer JD different from a data scientist JD?

The verbs change. A data engineer “builds,” “operates,” and “maintains” pipelines and infrastructure; a data scientist “models,” “experiments,” and “predicts.” If your bullets are full of pipelines, orchestration, and warehouse modeling, you are writing a data engineer posting. If they lean on statistics, experimentation, and machine learning, that is a data scientist, and a different person with different comp. Write one or the other. The role that tries to be both attracts generalists and loses the specialists you actually wanted.

Next Steps

Copy the template. Fill in the brackets with your real stack, your real data, and your real band. Delete the role types you are not hiring, and cut every “required” line that is actually a wish. Post it.

If you would rather have a recruiter pressure-test the JD, calibrate the salary range for your metro, or just put qualified data engineers in front of you without the 200-applicant slog, talk to one of our recruiters. We fill data and engineering roles in 17 days on average across 30-plus U.S. metros, through contract, contract-to-hire, and direct hire. When you are ready to budget the whole hire, the guide on the full cost of a data engineer hire and the senior salary guide pick up where this leaves off.