Level 5 Data Engineer Apprenticeship (ST1386) UK

By James Cotton, Founder, iO-Sphere

Most guides to this apprenticeship hand you a level description, an entry-requirements list, and no way to actually decide. That's the wrong shape for the real question. Levels aren't school years you climb from the bottom. They're entry points matched to the complexity of the work a role already holds. An experienced BI developer or software engineer often enters straight at Level 5 because their job already demands it. Someone whose role touches spreadsheets and dashboards is usually better served at Level 3 or 4.

So before the standard, the funding or the assessment, ask the practical question. And it has a number attached, which most guides never give you.

Can this role hold 418 hours of Level 5 engineering work?

Since August 2025 every apprenticeship standard publishes its own minimum off-the-job hours, replacing the old "20% of contracted hours" proxy. ST1386 sets 418, the highest of any standard we deliver.

Those hours are not classroom time. They are protected working hours the apprentice spends doing Level 5 engineering, and they have to come out of a real job. Providers schedule them differently: we take the standard's hours and smooth them across delivery, so across 15 months of training that lands at roughly six and a half hours a week.

So the question is not whether someone is capable of Level 5. It is whether the role contains 418 hours of it. Ask the line manager to name that work before anyone enrols: not the topics the apprentice will study, the work they will do. The next section gives you the test for reading their answer.

This is the check most shortlists skip, and it is why the wrong level is the main way a funded programme quietly fails: the learning detaches from the day job and becomes classroom theory with extra admin. 418 hours is a long time to spend discovering that.

How to read the answer you get

Here is the committed test. Level 5 fits if the role involves at least two of: owning a production pipeline end-to-end; choosing and operating a data storage or processing platform; writing and maintaining transformation logic in Python or SQL at a scale that breaks naive approaches; or mentoring others on data infrastructure decisions.

It does not fit if the role's outputs are reports, dashboards or ad-hoc queries, however sophisticated, because those outputs are downstream of the engineering work, not the engineering work itself.

The error we see most is employers confusing a technically fluent analyst (someone who writes Python to interrogate data) with a data engineer, who builds the system that makes the data available to interrogate. They are different jobs, and the apprenticeship can only certify the second. The tell-tale signal of a role that can't hold Level 5 work: when you ask the line manager to list the pipeline-building tasks they'll assign, the list turns into a list of questions they want answered from data. That is analysis, not engineering.

If the answer is no, here is where to go instead

Most guides bury this or skip it, because it sends readers away. It is the most useful section on the page: a wrong-level start costs a year, and the funding cannot be spent twice.

If the role genuinely doesn't contain data-engineering work (no pipelines to build, no platform to maintain, no data to move at scale), Level 5 is the wrong choice however keen the person is. You would be funding a syllabus the job can't feed.

Three routes are genuinely open instead, and it is worth naming where each one goes rather than gesturing at "a lower level".

The role's outputs are reports, dashboards or analytical findings. That is the Level 4 Data Analyst standard (ST0118), which develops people who work with existing data to answer business questions: querying, visualising, communicating insight. It carries 370 off-the-job hours rather than 418. Ours is the Level 4 Data Analyst apprenticeship.
The work is foundational data-handling, closer to preparing and maintaining data than building the systems that move it. That is Level 3, the Data Technician standard (ST0795), at 348 hours. Ours is Data & AI Essentials, and the Level 3 route explained in full covers who it suits.
Nobody in the role does data engineering at all yet. Then no apprenticeship is the answer this year. Build the work first, or start someone at the level the job holds now and revisit. The employer guide walks through both.

And if the ambition is statistical modelling and machine learning rather than building the pipelines that feed it, that's the data scientist path: a different discipline, one iO-Sphere doesn't deliver at this level, so weigh it on its own merits elsewhere. The test is always the work, not the ambition. Talk to whoever manages the role about what tasks they'll actually assign at Level 5 before anyone enrols. The syllabus can't supply what the role doesn't contain.

If the answer is yes, here is what you are committing to

Everything below assumes the role holds the work. If it does not, none of it rescues the decision.

The standard, and what it develops

The Level 5 Data Engineer apprenticeship is a work-based route that develops someone into a data engineer: the person who builds and maintains the pipelines and platforms that collect, move, clean and store data so analysts, dashboards and AI systems can use it. It runs on the Skills England-approved Data Engineer standard (ST1386), which sets out the knowledge, skills and behaviours the role requires.

Skills England maintains apprenticeship standards in England; it replaced the Institute for Apprenticeships and Technical Education (IfATE) on 2 June 2025. Funding policy now sits with the Department for Work and Pensions (DWP), following the machinery-of-government move confirmed on 16 September 2025. ST1386 is approved for delivery now, and a revised assessment plan replaces the current version for new starts on 18 December 2026 (Skills England, accessed 2026-07-02). Check the live standard page if you're planning a start around that changeover.

Level 5 sits above A-level and below a full bachelor's degree. On the domain ladder, it's the step beyond the Level 4 Data Analyst role: an analyst answers business questions with data; a data engineer builds the plumbing underneath so the data is there to answer them. If you're weighing those two, our data engineer vs data analyst comparison draws the line in detail.

What the apprentice needs before starting

Entry requirements are set by the training provider, not by the standard: Skills England publishes none for ST1386, so they vary. Typically you need a Level 3 qualification in a relevant area (any grade): two A levels in a similar subject, a Level 3 apprenticeship, a BTEC Extended Diploma, or genuine programming experience such as working knowledge of Python, or 2 to 3 years in a related role. Ask any provider for theirs. "No degree needed" is true. "No prior knowledge needed" is not. Level 5 is engineering work, and it assumes a real technical starting point.

Apprentices also need English and maths to Level 2 (roughly a GCSE grade 9 to 4 standard pass, or an Ofqual-regulated Functional Skills Level 2). For anyone aged 19 or over at the start, that exit requirement has been optional since 11 February 2025, though it can still be funded if they lack it and the employer agrees. For 16 to 18-year-olds it remains a completion condition. Existing employees are eligible, and this is often the most powerful use of the route: upskilling someone who already knows your systems.

What is covered, and how it is assessed

ST1386 covers production-grade data engineering: designing and building data pipelines, working with languages such as Python and SQL, managing data storage and processing, applying data governance and security, and doing it all to the reliability standard real systems demand. iO-Sphere teaches these on real data rather than toy examples. Apprentices work in Prism, a simulated e-commerce company built on 500M+ rows, so they practise pipeline-building in a safe sandbox before it counts on live systems.

Assessment is where this route parts company with an academic course. There is no written exam. The end-point assessment (EPA) has two methods, both built on the apprentice's real work: a project evaluation report with a presentation and questions, and a professional discussion. Each method is graded fail, pass or distinction, and the two combine into an overall grade of fail, pass, merit or distinction (ST1386 assessment plan, Skills England).

It is delivered by an independent assessment organisation, which the employer selects from the register of end-point assessment organisations. For anyone who dreads exams, that's a genuine relief: you're judged on what you can build, not what you can recall under a clock. Our end-point assessment glossary entry explains the mechanics.

How long it takes, and how the hours are protected

iO-Sphere delivers the Level 5 Data Engineer apprenticeship in 15 months of training plus a 3-month end-point assessment. The standard's typical duration is 24 months, and providers deliver it in different lengths (cambridgespark.com advertises 14 months, 2026-07-13), which tells you duration varies with how a provider structures delivery, not with the standard itself. Learning happens off-the-job during paid working hours: protected time, never "catch up at home."

What it costs

Any start you plan now lands in the 2026-27 funding year (starts from 1 August 2026). How the levy pays for it:

Levy-paying employer (pay bill over £3m). You draw on your apprenticeship service account, which covers training and EPA up to the funding band. The maximum government funding for ST1386 is £19,000 (Skills England). Confirm the live figure in the DWP apprenticeship funding rules before committing budget, as bands are revised periodically.
Non-levy employer (pay bill under £3m). You don't pay the levy but the apprenticeship is still funded, through government co-investment (£0 for an apprentice aged 16 to 24 at the start; 5% at 25+, with government paying 95%) or a levy transfer from a larger employer covering up to 100% of the cost. "We don't pay in" is not "we can't benefit."

If a levy account is exhausted, the 2026-27 rules still fund an apprentice aged 16 to 24 at the start at 100%; for an apprentice aged 25 and over the split becomes 75% government, 25% employer (v3, 29 July 2026). Level 5 is unaffected by the Level 7 funding withdrawal, which only hits master's-level starts aged 22 and over.

A note on apprenticeship units, and our interest in it. Short standalone units became levy-fundable on 28 April 2026, and you will see them offered as the flexible alternative. We do not deliver them, so read what follows knowing that. Our view is about the economics rather than anyone's intent: the funding per unit is small, the format is short, and a few dozen hours ends before the practice meets the edge cases where the learning actually lives.

For some organisations a unit is genuinely the right thing, when it is fully funded and nothing else is affordable, or as a bridge while a fuller route is planned. It is a poor destination. Judge one the way you would judge any training: by what the learner will have done by the end. On this page that test is 418 hours against a few dozen. The employer funding page works through the routes for your situation.

Where it leads

A Level 5 Data Engineer apprenticeship qualifies someone for data engineer roles and is a strong platform for progression toward senior engineering and, later, machine-learning engineering (a Level 6 standard, ST1398, exists). Data engineering is consistently among the better-paid data roles because the skills are scarce and the work is production-critical.

No single salary figure would serve you here: pay varies widely by sector, region and stack, and a stale point estimate does you no favours. Check a current, dated source such as the ONS ASHE tables or a live recruitment salary guide for the range that fits your market.

As a shape: apprentices earn a real salary throughout (the apprentice National Minimum Wage is £8.00/hour from 1 April 2026, the legal floor rather than the going rate, and most data-engineering employers pay well above it); qualified data engineers command materially more, and experienced engineers more again. The durable signal isn't any one quarter's figure. It's the structural, growing demand for people who can make data usable.

What to ask us, and everyone else on your list

Sort by four things, in this order:

What will the learner actually be doing? If the answer is "watching lectures", stop. If it's "building pipelines on real data", continue.
Who coaches? People who've done the job, or people who teach the subject? Ask directly.
Does the role hold Level 5 work? Confirm with the line manager what tasks they'll assign, the check most shortlists skip.
What does it run on? It should be ST1386, at Level 5. Then, and only then, fit the funding to it.

Notice funding comes last. "Is it free?" is the wrong first question. Decide the outcome first (what you want this person to be able to do, and how you'll measure it), then judge every route as if you were paying cash, and only then work out the funding. A funded course that doesn't change what someone can do is the most expensive training there is, because it spends the one resource you can't get back: working time.

Here's the position we'll stake plainly: capability, not technology, is the bottleneck, and delivery model is where capability is won or lost. The evidence is stark. MIT's NANDA initiative found 95% of GenAI pilots delivered no measurable P&L return despite $30 to 40bn of enterprise investment, naming learning, not infrastructure, as the core barrier ("The GenAI Divide", fieldwork Jan to Jun 2025). RAND found more than 80% of AI projects fail, twice the rate of non-AI IT projects, the top root cause being a misunderstanding of the problem to be solved (RAND RRA2680-1, August 2024). Neither is a tooling story. Both are capability stories.

That capability gap explains why the format matters, not just the provider. A bootcamp or short course can transfer knowledge; it can't transfer the habit of applying it under real operational pressure, on real systems, over months. The apprenticeship model forces that repetition: off-the-job time is protected working time, not a side project. The EPA then certifies the output, not the attendance. That is why the delivery model isn't a differentiator between providers alone. It's the structural reason the apprenticeship format outperforms shorter interventions for production-grade capability.

So when you apply the four comparison criteria above, delivery model is the thing they're all testing for. Lectures and theory-first modules produce people who can describe a data pipeline. Building real pipelines, coached by people who've done the job, produces people who can build one.

Our own answers to those four, so you can hold everyone else to the same test. What will the learner be doing: building pipelines, first in Prism on 500M+ rows of real data, then on your systems. Who coaches: practitioners who have done the job, not academics, and you can ask for their backgrounds. Does the role hold Level 5 work: we will not enrol someone whose role does not, because a wrong-level start wastes your funding and their year. What it runs on: ST1386 at Level 5, with the 418 hours protected inside working time.

FAQ

What level is the Data Engineer apprenticeship?

The Data Engineer apprenticeship is Level 5, run on the Skills England-approved standard ST1386. Level 5 sits above A-level and below a full bachelor's degree: technical engineering work, a step beyond the Level 4 Data Analyst role.

How long does the Level 5 Data Engineer apprenticeship take?

The standard's typical duration is 24 months. iO-Sphere delivers it in 15 months of training plus a 3-month end-point assessment. Other providers advertise shorter timelines (cambridgespark.com lists 14 months, 2026-07-13), so duration depends on how a provider structures delivery.

What are the entry requirements?

Entry requirements are set by the provider rather than the standard, since Skills England publishes none for ST1386. Typically you need a Level 3 qualification in a relevant area in any grade: two A levels in a similar subject, a Level 3 apprenticeship, or a BTEC Extended Diploma, or equivalent such as working programming experience (for example Python) or 2 to 3 years in a related role. A degree isn't required, but genuine technical grounding is. Apprentices also need English and maths to Level 2, though for those aged 19+ that exit requirement has been optional since 11 February 2025.

How is the Level 5 Data Engineer apprenticeship funded?

The maximum government funding for ST1386 is £19,000 (Skills England); confirm the live figure before budgeting, as bands are revised. Funding runs through the apprenticeship levy. A levy-paying employer (pay bill over £3m) draws on their apprenticeship service account; from 1 August 2026, if the account is exhausted, they co-invest 25% (75% government) for apprentices aged 25 and over; a 16 to 24 year old at the start remains 100% funded. A non-levy employer (pay bill under £3m) pays 5% under co-investment, or £0 if a larger employer transfers levy funds covering the cost.

Is there an exam?

No written exam. The end-point assessment has two methods, both built on the apprentice's real work: a project evaluation report with a presentation and questions, and a professional discussion. Each is graded fail, pass or distinction, and together they give an overall grade of fail, pass, merit or distinction (ST1386 assessment plan, Skills England). It is delivered by an independent assessment organisation. You're assessed on what you can build, not on recall under a clock.

Should I pick Level 5, or a lower level?

Pick the level whose work genuinely exists in the role now. Level 5 fits if the role involves at least two of: owning a production pipeline end-to-end, running a storage or processing platform, maintaining Python or SQL transformation logic at scale, or mentoring on data infrastructure. Reports, dashboards and ad-hoc queries sit downstream of the engineering, and a fluent analyst is not yet an engineer: a role built on them points to the Level 4 Data Analyst standard (ST0118), or to Level 3 for foundational data-handling.

Can an existing employee do the Level 5 Data Engineer apprenticeship?

Yes. Existing employees are among the most effective uses of this route. If someone is already doing parts of data-engineering work but lacks the formal grounding, the apprenticeship formalises and deepens what they can do on real systems they already know. The eligibility rules are the same: Level 3 prior qualification or equivalent experience, English and maths to Level 2 (optional exit for those aged 19+ since February 2025), and a role that genuinely contains Level 5 engineering work.

If the Level 5 work genuinely lives in the role, the next step is a conversation about what your team will build, not a brochure. Explore our Level 5 Data Engineering apprenticeship or talk to us about whether it's the right level for the work you have. →