How to Hire a Database Administrator

Hiring a database administrator means finding someone who can keep your production data alive, recoverable, and fast under real load. Budget $85,000 to $165,000 depending on the platform and seniority, expect the search to take 30 to 60 days for a mid-level candidate and longer for senior Oracle or cloud-native specialists, and know before you write the job description that “DBA” describes at least four different jobs. Posting the wrong one is the single most expensive mistake in this search.

Last fall a client in Orange County called us about a DBA hire that had been open for 11 weeks. Good company, funded role, reasonable salary. The problem was obvious from the JD. They wanted someone to manage a production SQL Server environment, tune queries on a 4TB analytics warehouse, run the AWS RDS migration they’d been planning for two quarters, and “help the dev team with schema design as needed.” That is three jobs. Maybe four. The production DBA who keeps your 99.99% uptime SLA is not the same person who redesigns your schema for a microservices migration. The candidate pool for each of those is different, the comp band is different, and the interview questions that actually work are different. We split the req into two roles, filled both inside 35 days, and the combined salary was only $12,000 more than the single inflated req they’d been failing to fill for three months.

Devin Hornick, partner at KORE1. I run our engineering and technical staffing practice, which means DBA searches land on my desk when they’ve already gone sideways once. Quick note on incentives before we go further: we’re a technical staffing firm and we get paid when you hire through us. I’ll flag the parts of this guide where you probably don’t need us. A few sections will make it clear when you might.

What a Database Administrator Actually Does

A database administrator keeps your data alive, recoverable, and performing under load. The short version of the job: install and configure database software, monitor the health of running systems, manage backups so disaster recovery actually works when you need it, control who can access what, and tune queries and indexes before the applications that depend on the data start timing out during peak traffic.

The day looks different depending on where you sit. A production DBA at a financial services firm? Their Tuesday morning is backup logs from the overnight window, checking replication lag between primary and standby, and then an hour figuring out why a stored procedure that normally finishes in 400 milliseconds took 12 seconds during the batch run. A development DBA at the same firm is in a different meeting entirely, arguing with the application team about whether their proposed table structure will survive the read patterns they haven’t thought about yet.

The Bureau of Labor Statistics reports a median annual wage of $104,620 for database administrators as of May 2024, with database architects earning a higher median of $135,980. Projected employment growth sits at 4% through 2034, roughly matching the national average, with about 7,800 openings annually. Those numbers are fine as context. They are not fine as a hiring budget, because the spread between a junior MySQL DBA and a senior Oracle RAC specialist is roughly $80,000, and the BLS median sits right in the middle where almost nobody actually works. Our DBA salary guide breaks the numbers down by platform, experience, and geography if you need the full picture before writing an offer.

Four Kinds of DBA (and Posting the Wrong One Wastes Two Months)

This is where most searches go wrong. A hiring manager writes a JD that says “database administrator,” LinkedIn serves it to 300 people, and 250 of them do a fundamentally different version of the job than what the team actually needs. The remaining 50 who are a fit see a JD that reads like four roles mashed together and move on.

Here’s the split.

Salary ranges above draw from our placement data, cross-referenced with Glassdoor, ZipRecruiter, and Built In as of early 2026.

The overlap between these roles is smaller than most people assume. A production DBA who has spent eight years keeping Oracle Data Guard failovers clean is not automatically going to be good at schema design for a greenfield microservices project. They might be. But the skills that made them excellent at operations, the paranoia about change management, the obsessive backup verification, the 3 AM composure, those are orthogonal to the skills that make a good development DBA. And a cloud DBA is a different animal entirely. The person managing your Aurora PostgreSQL clusters is spending half their day in the AWS console and the other half writing infrastructure-as-code. Their SQL skills might be strong, or they might be adequate. What they’re really good at is cost modeling and automation.

Pick one for the req. If you genuinely need two, post two.

DBA Salary Ranges in 2026

The BLS median of $104,620 is a starting point but it obscures the real variance. Platform specialization, cloud skills, and geography create a $80,000+ spread within the same job title.

Oracle still commands the highest base among traditional on-prem platforms. Makes sense. The licensing is complex, the tooling is proprietary, and companies running Oracle RAC or Exadata environments tend to be large enterprises with real budgets. Cloud DBA roles carry a $15,000 to $30,000 premium over equivalent on-prem experience because the skill set is genuinely different, not because the title sounds fancier.

Geography hasn’t gone away, even if people keep saying it has. Bay Area and New York add 12 to 20 percent. Texas, the Southeast, and the Midwest subtract 8 to 15 percent. Remote positions have compressed some of that gap but not all of it. A fully remote senior PostgreSQL DBA in Atlanta is making $130K. The same person in San Francisco, also remote, is making $150K. Same company, same role, different zip code on the W-4. Companies that refuse to adjust for geography lose candidates to companies that will.

For the full breakdown by platform, city, and contract rates, see our Database Administrator Salary Guide.

The Skills That Separate a Real DBA From a Resume

Every DBA job posting lists SQL, backups, performance tuning, and security. Every applicant claims proficiency in all four. The hiring manager who screens on those keywords alone will interview fifteen people and hire the wrong one.

What actually matters depends on which of the four DBA types you’re hiring, but there’s a core layer that applies everywhere.

Platform-Specific Skills (Pick Your Stack)

SQL Server. T-SQL fluency. SSMS is table stakes. What you’re really screening for: can they read an execution plan, identify a missing index, and explain why the query optimizer chose a table scan instead of a seek? Ask them to walk through a real slow-query investigation. If their answer starts with “I would check the indexes” and stops there, they’re junior regardless of what their resume says. Look for experience with Always On Availability Groups, log shipping, and SSIS if you do ETL work.

Oracle is its own ecosystem. PL/SQL, RMAN, Data Guard, ASM, RAC if you’re running clustered. The Oracle DBA talent pool just expanded significantly because of the March 2026 layoffs that put roughly 30,000 Oracle employees on the market, including experienced DBAs who ran production RAC environments at Oracle itself. We wrote a full breakdown of how to hire Oracle DBAs in the current market if that’s your platform.

PostgreSQL. The fastest-growing database in the world right now. The 2025 Stack Overflow Developer Survey showed PostgreSQL at 55.6% adoption among professional developers, up from 48.7% the year before. Biggest single-year jump the survey has ever recorded. What this means for hiring: the PostgreSQL DBA candidate pool is growing, but much of it is developers who know PostgreSQL well, not dedicated DBAs who’ve managed it in production at scale. Screen for pg_dump vs. pg_basebackup understanding, replication slot management, vacuum tuning, and connection pooling with PgBouncer. If they’ve only ever used PostgreSQL through an ORM, that’s a developer, not a DBA.

Cloud platforms require a different mental model. The DBA managing your RDS or Azure SQL Managed Instance is not manually configuring storage or patching the OS. They’re making decisions about instance classes, read replicas, automated backups retention, parameter groups, and cost. A senior cloud DBA should be comfortable with Terraform or CloudFormation, understand VPC networking enough to configure security groups and private endpoints, and be able to explain why a db.r6g.2xlarge is or isn’t the right choice for your workload. The Gartner DBMS forecast projects the overall database management market growing 18.4% in 2026, with cloud-native expansion as the primary driver. The hiring implications are clear: pure on-prem DBA skills are still needed, but the candidates who can work across both environments command the premium.

The Soft Skills Nobody Puts in the JD but Everybody Needs

Communication. Not in the generic “good communicator” sense. Specifically: can this person explain to a non-technical product manager why the schema change they want will take three weeks instead of three days, without making the PM feel stupid, and without caving when the PM pushes back? We’ve placed DBAs who were technically flawless and got fired inside six months because they couldn’t have that conversation without escalating it into a political problem.

On-call composure. A production DBA who panics during an outage creates two problems. Ask candidates to describe a real incident they handled. What broke, what they did first, what they would do differently. The answer reveals more than any whiteboard exercise.

Documentation discipline. The DBA who doesn’t document the recovery procedure is the DBA whose departure costs you $50,000 in institutional knowledge reconstruction. Not a nice-to-have.

Certifications: When They Matter and When They Don’t

Oracle Certified Professional (OCP) matters if you’re running Oracle. It’s a genuinely difficult certification that requires real knowledge. The Microsoft Azure Database Administrator Associate (DP-300) matters if you’re moving to Azure SQL. Both carry a 10 to 20 percent salary premium in our data.

Generic SQL certifications from online course platforms? Almost useless as a hiring signal. They test textbook knowledge, not production judgment. A candidate with no certifications and six years of production PostgreSQL experience beats a candidate with three certifications and no production incidents on their record. Every time.

Where the DBA Candidate Pool Actually Is Right Now

Three things shifted the DBA talent market in the last 12 months.

The Oracle layoffs in March 2026. Up to 30,000 employees cut in a single round. That number included experienced DBAs who ran some of the highest-transaction-volume Oracle environments on the planet. If you need Oracle talent specifically, this is the best hiring window in a decade. It will close. These people are getting snapped up. We’ve already placed a dozen from that pool.

PostgreSQL’s adoption surge. When a database goes from 48.7% to 55.6% developer adoption in a single year, the DBA hiring pipeline follows about 18 months behind. Right now you’re in the gap. Lots of companies want PostgreSQL DBAs. The dedicated production DBA pool for PostgreSQL is still thin compared to SQL Server or Oracle. The workaround: hire strong DBAs from adjacent platforms who are genuinely motivated to switch. An experienced SQL Server DBA can become productive on PostgreSQL in 60 to 90 days. An Oracle DBA, closer to 90 to 120. The core skills transfer. The tooling doesn’t, but it can be learned.

Cloud migration everywhere. Gartner’s 2026 forecast pegs the DBMS market at $161 billion, with cloud-native expansion as the biggest growth driver. Companies that ran everything on-prem two years ago now have a foot in AWS or Azure and need someone who can manage both. That hybrid DBA, the person comfortable with on-prem SQL Server and also fluent in RDS or Azure SQL Managed Instance, is the hardest profile to fill and the most in demand.

Writing a DBA Job Description That Works

The job descriptions that sit open for 90 days share three patterns. They ask for everything. They specify nothing. And they bury the comp.

Be specific about the platform. “Experience with relational databases” tells the candidate nothing. “Primary responsibility: production SQL Server 2019 environment, 12 databases, largest at 2.1TB, Always On AG across two data centers” tells them exactly what they’re walking into. Good candidates self-select on specificity. Bad candidates self-select on vagueness.

State the on-call expectation. If the role carries a pager, say so in the first three bullet points, not buried in a paragraph at the bottom. DBAs who are willing to do on-call want to know the rotation (one week in four? one in six?) and whether there’s a comp differential. DBAs who won’t do on-call want to know that before they spend 45 minutes on a phone screen. Either one is perfectly reasonable. Hiding it is not.

Post the salary range. In 2026, fourteen states plus DC require salary transparency in job postings. Even where it’s not legally required, posting the range cuts your time-to-fill. We track this internally. Reqs with posted ranges fill 11 days faster on average across our IT staffing practice. The candidates who apply are better matched because the ones outside your budget never wasted your time.

Drop the degree requirement unless you actually mean it. The BLS notes that most DBA positions “typically” require a bachelor’s degree. In practice, roughly a third of the DBAs we’ve placed in the last two years didn’t have a four-year degree. They had certifications, years of hands-on production experience, and the kind of troubleshooting instinct that comes from managing systems under real pressure. If your HR system auto-rejects candidates without a BS in Computer Science, you’re filtering out people who might be your best hire.

Interview Questions That Actually Tell You Something

Skip the SQL trivia. A candidate who can recite the difference between INNER JOIN and LEFT JOIN from memory is not necessarily the candidate who can figure out why your nightly ETL job just failed at 2:14 AM on a Saturday. The questions that work are scenario-based, platform-specific, and designed to reveal judgment, not recall.

For Production DBAs

“It’s 2 AM and you get paged. The primary database server is responding but queries are timing out after 30 seconds. Walk me through your first ten minutes.” What you’re listening for: do they check replication status, disk I/O, blocking chains, and recent deployments in a logical order? Do they know what to check before they start changing things? The DBA who immediately says “I’d restart the service” is telling you something important about their risk tolerance.

“Your most recent full backup completed successfully but the backup file is 40% smaller than last week’s. Nobody made structural changes. What do you investigate?” Good answer territory: data deletion, compression changes, backup corruption that reports success but isn’t, statistics that changed the backup strategy. Great answers include: “I’d do a test restore before I investigate anything else, because if that backup is bad, finding out why it’s small matters less than knowing whether my recovery point is gone.”

For Development DBAs

“A developer submits a pull request that adds a new index on a 500-million-row table in production. The index creation will lock the table. How do you handle this?” You want the candidate who asks questions back. What engine? What version? Is online index creation an option? What’s the maintenance window? Can we build it on the replica first? The candidate who says “I’d approve it” or “I’d reject it” without asking for context is not thinking at the right level.

For Cloud DBAs

“Your Aurora PostgreSQL cluster’s monthly bill jumped 35% and nobody changed the application code. Where do you look?” Strong candidates talk about read replica utilization, storage I/O charges, snapshot retention bloat, instance class changes by another team, and cross-region replication costs. The best answer I’ve heard in an actual interview started with “I’d pull the Cost Explorer breakdown by usage type before I touched anything else, because the bill tells you where the money went before you start guessing.”

Red Flags in DBA Interviews

Can’t describe a real incident. Not a hypothetical, a real one. Every working DBA has a war story. If the only incidents they can describe are from training labs or certification study guides, they haven’t managed production systems under pressure.

Blames the application team for everything. Sometimes the application code is the problem. But the DBA who defaults to “the developers wrote bad queries” as the root cause for every performance issue is going to create an adversarial relationship that slows your entire team down.

No questions about the environment. A serious DBA candidate will ask about the database platform, version, size, replication topology, backup strategy, and who they report to. Silence there is a signal. Either they haven’t managed a real environment or they don’t care enough about yours to ask.

Contract vs. Direct Hire for DBA Roles

Both models work. The right one depends on why you’re hiring.

Direct hire makes sense when the role is permanent, the database environment is stable, and you want someone who will accumulate institutional knowledge over years. Most production DBA roles should be direct hire because the person managing your backups and disaster recovery needs to know the environment deeply enough to act fast when things break. The placement fee (18 to 25 percent of first-year salary) is a one-time cost that pays for itself the first time your DBA prevents a data loss event that would have cost you twenty times that number.

Contract staffing makes sense for migrations, platform upgrades, specific projects with defined timelines, and covering gaps while you search for a permanent hire. A six-month contract to migrate from on-prem SQL Server to Aurora PostgreSQL is a common engagement we run. The markup runs 30 to 45 percent over the contractor’s pay rate, billed weekly, with conversion to direct hire available at a negotiated buyout if the person is a fit and the project becomes ongoing.

Contract-to-hire is the most common model we see for DBA roles specifically. The hiring manager wants to see how the candidate handles a real incident before committing. Makes sense. A 90-day trial on contract, converting to permanent if it works, gives both sides an out. About 70% of our DBA contract-to-hire engagements convert.

What Hiring Managers Keep Asking Us

So what’s the realistic timeline to fill a DBA role?

30 to 45 days for a mid-level SQL Server or MySQL DBA in a major metro. 45 to 75 days for senior Oracle or cloud-native specialists. Faster if you post the salary range and slower if you don’t. That 11-week open req I mentioned at the top? Could have been 5 weeks with a clearer JD and realistic comp expectations.

Do we actually need a full-time DBA or can we outsource it?

Depends on the database count and criticality. Under five databases with no high-availability requirement? A managed DBA service or part-time contractor can handle it. Over ten production databases, or anything with an SLA that involves the word “nines”? You need a person, not a service. The in-between is where the real judgment call lives, and it usually comes down to how fast you need someone to respond when something breaks at an inconvenient time.

Should we hire a DBA or just train a developer?

Short answer: it depends on what’s breaking. A developer who writes good SQL and understands indexing can handle schema design and query optimization for a small application. That same developer will not be the person you want managing backup rotations, testing disaster recovery procedures, or handling a replication failure on a Sunday afternoon. Those are operational skills built over years of incidents, not a weekend workshop. If your database environment is growing past the point where “the dev team handles it” feels safe, it’s past the point where you need a DBA.

Oracle or PostgreSQL for a new project in 2026?

Wrong question, slightly. The database platform decision should drive the DBA hire, not the other way around. But since people ask: PostgreSQL adoption hit 55.6% among professional developers in the 2025 Stack Overflow survey, the candidate pool is expanding, and the licensing cost is zero. Oracle makes sense when you’re already in the Oracle ecosystem, when the application requires features that PostgreSQL doesn’t match (Real Application Clusters, Exadata integration, advanced partitioning at multi-petabyte scale), or when the team already has deep Oracle expertise. For most new projects that don’t hit those criteria, PostgreSQL is where the momentum is.

How do you tell a real DBA from someone who just lists SQL on their resume?

Ask them what went wrong. Not hypothetically. What actually broke, in production, on their watch, and what they did about it. The real DBA lights up. They’ll tell you about the time a developer dropped a production table on a Thursday afternoon, or the backup that reported success for six weeks but was actually writing to a full disk, or the replication lag that nobody noticed until a customer saw stale data. The person who just lists SQL on their resume will give you a textbook answer about best practices. The difference is impossible to miss once you know what to listen for.

Is the DBA role going away because of cloud and automation?

No. The role is changing. A 2026 analysis from Database Trends and Applications put it well: DBA skills will continue to be in demand, but they’ll be “recast as one facet of the role” alongside cloud platform management, automation, and infrastructure-as-code. The routine monitoring and patching work that used to consume 40% of a DBA’s week is increasingly automated. What isn’t automated is the judgment: deciding when to fail over, diagnosing performance problems that don’t match documented patterns, designing schemas that will scale, and making the call at 2 AM about whether to wake up the VP of Engineering. That’s the job now. It pays more than the old version, not less.

Start the Search

If you’ve read this far and the DBA search still feels manageable, run it yourself. Seriously. Strong internal recruiting teams fill DBA roles every day, and the framework above should get you most of the way there.

If the req has been open for more than six weeks, or the platform is specialized (Oracle RAC, Aurora PostgreSQL at scale, multi-cloud), or you’re not sure which of the four DBA types you actually need, that’s where we tend to get the call. Reach out to our technical staffing team and we’ll start with an intake call to figure out whether this is a search we can actually help with. Sometimes the answer is “you don’t need us,” and we’ll say that directly.