Free job simulation · AI and data
AI and LLM application engineer job simulation: build an assistant, evaluate it, handle an injection incident
An LLM engineer job simulation is a hands-on exercise that mirrors the role. For a simulated housing association you design chunking and retrieval for a staff policy assistant, run an evaluation set, catch a prompt change that makes it invent answers and respond when a hidden instruction turns up in an indexed document. It is practice, not employment, and it ends with a release pack.
You join Halden Applied AI Ltd (simulated company), a ai application studio in Bristol. Days 1 to 5 are free, with about 10 h of tickets. Premium continues to day 20. It is practice, not employment.
Days 1 to 5 are free. You need a free account and no card. Create an account.
- Free days
- 1 to 5
- Free tickets
- about 10 h
- Project documents
- 8
- Certificate, optional
- £14.99
The project
You join Halden Applied AI Ltd (simulated company)
Halden Applied AI Ltd (simulated company) is a ai application studio in Bristol. You help its team deliver a project for Northfield Housing Association Ltd, at Ask Northfield, a staff assistant for HR, IT, finance and safety policies (ticket LLM-031).
You are the junior AI engineer on Ask Northfield, an assistant that answers staff questions from 25 policy documents, with citations. The Lead AI Engineer gives you the requirements, a design note with mistakes in it, a golden question set, two evaluation runs and a trace log. Over four weeks you decide how to chunk the documents, tune retrieval in a lab, find that a friendlier prompt has started inventing answers, work out why the pilot cost doubled, check whether the judge model can be trusted, handle a change request for a scanned archive, find the hidden instruction that told two users to email their password, count the personal data in the logs and decide what to say when the client wants to launch before the numbers are ready. Some pressure comes from senior people. Your job is to be useful and truthful at the same time.
What you will do
- Decide chunking and retrieval with counts, and tune them in an EDWartens Lab exercise
- Run an evaluation set, read the run table and catch a prompt change that makes the assistant invent answers
- Work out cost and latency from a trace log and find what doubled them
- Find a hidden instruction in a scanned document, write the incident report and fix the process
- Count personal data in logs, protect access in code and write a model and system card
The people you deal with
- Head of Engineeringcompany
- Lead AI Engineercompany
- Delivery Managercompany
- Security Reviewercompany
- Client Project Managerclient
- Client Head of HRclient
- Client IT Managerclient
- Client Data Protection Officerclient
The company, client and site are fictional. The people are fictional too.
Look inside
What the simulator looks like
The workplace, your tickets, the documents and the team chat, as a learner sees them.
Start in the workplace, with the team around you. 
Today: the morning email and your tickets for the day. 
Work: every ticket, by day, with its status. 
A ticket, written the way a colleague would write it. 
A project document with its number, revision and status. 
Ask a colleague. They are AI colleagues: they help you think, they do not hand out answers.
Screenshots from the simulator. Company, people and data are fictional. Portraits are AI-generated.
Meet the team
The colleagues you work with
You can read their messages and ask them questions. They are AI colleagues, clearly labelled. They help you think a ticket through; they do not hand out its answer.
- CRCormac RaffertyHead of EngineeringSimulated colleague, company side
- TMThandiwe MokoenaLead AI EngineerSimulated colleague, company side
- HRHollie RamsdenDelivery ManagerSimulated colleague, company side
- STSamir TahirSecurity ReviewerSimulated colleague, company side
- GUGwen UnderwoodClient Project ManagerSimulated colleague, client side
- CTClive ThwaiteClient Head of HRSimulated colleague, client side
- MVMandeep VirdiClient IT ManagerSimulated colleague, client side
- PWPhilippa WrayClient Data Protection OfficerSimulated colleague, client side
Fictional people. Portraits are AI-generated.
What you get
What you walk away with
Help build and test a document-grounded assistant: chunking, retrieval, evaluation, safety and a release pack.
What you produce
Classification of sample staff questions
Requirements table
Review comments on the chunking and retrieval design note
Design review
Evaluation table for two runs
Evaluation report
Bug report: the assistant invents answers
Bug report
Impact note for a change request
Change controlPremium
Incident report: hidden instruction in a document
Incident reportPremium
Model and system card
DocumentationPremium
Release pack for Ask Northfield
CapstonePremium
Skills you practise
- Evaluation and metrics
- LLM safety and prompt injection
- Written communication
- Retrieval and RAG
- Analysis and statistics
- Data protection and privacy
- AI governance and documentation
- Reading specifications
Roles to look for
- AI Engineer
- LLM Application Engineer
- Generative AI Developer
- Junior AI Engineer
101 live openings on the UK jobs board, checked 24 Oct 2026. See the jobs
Your proof
- A certificate anyone can verify. Optional, and included in Premium.
- An entry on your EDWartens profile under Work experience, marked as a job simulation.
- A work record with your skill scores, with Premium.
Free courses that go with it
What you can talk about at interview
- You can explain how you chunk documents, tune retrieval and check the assistant against a golden question set.
- You can walk through two evaluation runs and say which one is better and why.
- In Premium, you can describe an incident where a document held a hidden instruction, and how you wrote up the cause and the fix.
Day by day
Days 1 to 5, ticket by ticket
Self-paced. A day opens when you hand in the previous day's tickets. Every ticket is checked, and you see the model answer after it.
Day 1
about 1 h 35 min- Induction: retrieval, hallucination, injection and loggingQuiz · 45 min
- Classify ten sample staff questionsTable · 50 min
Day 2
about 2 h 40 min- Count the chunks and size the indexCalculation · 1 h 10 min
- Lab: tune chunking, retrieval and the refusal thresholdHand-in · 1 h 30 min
Day 3
about 1 h 55 min- Review the chunking and retrieval design noteDocument review · 1 h 20 min
- HR wants the pay spreadsheet in the assistantDecision · 35 min
Day 4
about 2 h 50 min- Lab: write a prompt to a specHand-in · 1 h 30 min
- Evaluate R-014 and R-015: the run tableCalculation · 1 h 20 min
Day 5
about 1 h- Bug report: the assistant invents answersWriting · 1 h
Premium: days 6 to 20about 22 h 30 min more
After day 5 the project goes on with a change request, site work, faults, a handover and a capstone reviewed by an engineer.
Day 6
about 1 h 30 min- Why did cost double? Cost and latency from the trace logCalculation · 1 h 30 min
Day 7
about 1 h 30 min- Calibrate the judge against the human checkCalculation · 1 h 30 min
Day 8
about 1 h 40 min- Impact note for change request CR-02Writing · 1 h 40 min
Day 9
about 1 h 20 min- Size the re-index for the scanned archiveCalculation · 1 h 20 min
Day 10
about 1 h 20 min- Retrieval recall and the old policy that keeps winningCalculation · 1 h 20 min
Day 11
about 1 h 30 min- Read the OCR text before it goes back in the indexDocument review · 1 h 30 min
Day 12
about 1 h 30 min- Lab: a tool-using flow that stays within limitsHand-in · 1 h 30 min
Day 13
about 1 h 20 min- Lab: find and remove the personal dataHand-in · 1 h 20 min
Day 14
about 1 h 20 min- How much personal data is in the trace log?Calculation · 1 h 20 min
Day 15
about 1 h 30 min- Incident report INC-07: the hidden instructionWriting · 1 h 30 min
Day 16
about 40 min- "We would like to go live on Monday"Decision · 40 min
Day 17
about 1 h 30 min- Write the model and system cardWriting · 1 h 30 min
Day 18
about 4 h- Capstone: the release pack for Ask NorthfieldHand-in · 4 h
Day 19
about 1 h- Release note and handover email to the clientWriting · 1 h
Day 20
about 50 min- A conversation about your own workViva · 50 min
Real documents
The documents you will work with
Each one has a number, a revision and an approver, the way a project document does. You read them, check them and write your own.
LLM-TKT-031
Ticket LLM-031: staff policy assistant that answers from the handbook with citations
Rev ATicket
LLM-REQ-001
Requirements and acceptance sheet
Rev ARequirements
LLM-STD-001
AI engineering team standards (one page)
Rev BStandard
LLM-COR-001
Document corpus inventory
Rev ACorpus inventory
LLM-GLD-001
Golden evaluation set
Rev AEvaluation set
LLM-PRM-001
Prompt registry: P-004 Ask Northfield system prompt
Rev APrompt registry
LLM-EVL-001
Evaluation method and run results R-014 and R-015
Rev AEvaluation report
LLM-DES-001
Chunking and retrieval design note
Rev ADesign note
Documents that arrive in Premiumchange requests, records, handover
LLM-CHG-002
Change request CR-02: scanned policy archive and effective-date filter
Rev AChange requestPremium
LLM-REQ-001
Requirements and acceptance sheet, with CR-02
Rev BRequirementsPremium
LLM-COR-001
Document corpus inventory, archive batch 1
Rev BCorpus inventoryPremium
LLM-PRM-001
Prompt registry: P-004 v4
Rev BPrompt registryPremium
LLM-TRC-001
Trace log and price list, weeks of 5 and 12 October
Rev ATrace logPremium
LLM-OCR-001
OCR extract of archive batch 1: IT-009 Laptop returns
Rev AOCR extractPremium
LLM-INC-001
Incident record INC-07: instruction hidden in an indexed document
Rev AIncident recordPremium
LLM-MCD-001
Model and system card: Ask Northfield
Rev AModel cardPremium
LLM-RBK-001
Runbook: Ask Northfield
Rev ARunbookPremium
LLM-HOV-001
Release and handover pack requirements
Rev AHandover indexPremium
Tools
In the industry, and here
| In the industry | Here |
|---|---|
| Notebook or scripts for evaluation and log analysis | Your own spreadsheet or Python on the downloaded run, trace and section files |
| Vector database and embedding service | The retrieval exercise in EDWartens Lab, and the chunk and index counts you work out yourself |
| Prompt registry and evaluation tooling | A numbered registry document and the golden set and run files |
| Tracing, ticket tracker and incident tooling | A trace log file, numbered tickets and an incident record |
Standards
The standards the work follows
- OWASP Top 10 for LLM applications
- ISO/IEC 42001
- UK GDPR
In the UK
The AI and LLM Application Engineer role in a UK setting
Who employs this kind of engineer: AI and software studios, consultancies, financial services, housing, health and public sector teams building assistants on their own documents.
Where: Roles are advertised across the UK, with larger groups in London, Manchester, Edinburgh, Cambridge, Bristol and Leeds, and many are hybrid or remote. This is context, not a statement about vacancies.
UK standards and regulations to know about
UK GDPR and the Data Protection Act 2018
The rules for personal data. The ICO publishes guidance, including on AI and data protection.
NCSC guidelines for secure AI system development
UK-led guidance on building AI systems securely.
OWASP Top 10 for LLM Applications
The common list of risks such as prompt injection.
UK regulators' AI principles and ISO/IEC 42001
The UK takes a principles-based approach through existing regulators. ISO/IEC 42001 is a management system standard for AI.
These are named so that you recognise them. A job simulation does not teach them as a qualification, the tasks use simplified data, and you should always read the current edition of anything you rely on.
Your proof, optional
A certificate you can verify
The free days are free. The graded, verifiable certificate is an optional extra, and it is included in Premium.
- A graded, verifiable certificate with your score and band
- A public verification page and QR code
- One-click add to LinkedIn as a licence or certification
- Issued by WARTENS LTD. It is a training certificate, not an accredited qualification, and the company is simulated
Optional: £14.99 including VAT, once you pass.
Job Simulation Certificate
Your name
completed the AI and LLM Application Engineer job simulation at Halden Applied AI Ltd (simulated company). Your score and band are shown on the real certificate.
- Number
- EDW-JS-0000
- Issued by
- WARTENS LTD
Not an accredited qualification. A verification page and QR code are printed on the real certificate.
Prepare
Free courses and labs that get you ready
Linked from the tickets that need them. You can read them before you start.
Free courses
- Generative AI and LLM Foundations
- RAG and Chatbots with LangChain
- AI Security and the OWASP Top 10 for LLMs
- Prompt Engineering for Engineers
- Python for AI and Engineering Data
- Private AI with Local LLMs: Ollama, Open WebUI and Offline RAG
- ISO 42001 and AI Governance
- AI Essentials for Work: ChatGPT, Claude, Gemini and Copilot
A course that goes with this role
AI Module
A paid, self-paced course in AI and machine learning for industry: 83 lessons, 5 graded projects and career sessions. It is a separate product: you do not need it to finish the job simulation, and it does not lead to a job. The certificate is a certificate of completion, not a regulated qualification.
- Price
- £2,688 including VAT (£2,240 + VAT); or start with a £100 deposit
- Format
- Self-paced online, 12 months' access
About the AI and LLM Application Engineer role
The incident ticket: a hidden instruction in an indexed document
Prompt injection is when input to a model changes its behaviour in ways the developer did not intend. In this project an instruction hidden in a scanned policy has been indexed, and the assistant may follow it.
You read the OCR text before it goes back in the index, write the incident report and fix the process that let it in. The case teaches the habit of treating retrieved text as untrusted input. A related task finds personal data in the trace logs and protects access in code.
The scorecard: why a demo is not evidence
A demo shows that an assistant can answer one question. An evaluation set shows how often it is right across many, and whether a change made it worse.
You work with a golden set and a run table, read two runs side by side and catch a prompt change that makes the assistant invent answers. You then calibrate an automated judge against a human check. The deliverable most job posts ask for, a documented evaluation, is the one you practise here.
Cost, latency and retrieval choices
Language model features have running costs. From a trace log you work out why cost doubled, and what changed in the retrieval or the prompt.
You also size an index from document counts, tune chunking, retrieval and a refusal threshold in an EDWartens Lab exercise, and check recall when an old policy keeps winning over a new one. All figures come from the files in the pack, so your answers are your own working.
Reading the numbers yourself is the point, because cost and latency questions in an interview are about reasoning, not tools.
A corpus of policies, and the documents you write
The assistant answers from staff policies on HR, IT, finance and safety. That is a realistic corpus: long, overlapping and sometimes out of date.
You write a bug report, an impact note for a change request, a model and system card and a release note. ISO/IEC 42001 and the OWASP Top 10 for LLM applications frame the governance and security thinking, and UK GDPR frames the data handling.
Writing for non-engineers, such as a release note for a client, is part of the job and is marked as such.
What to learn next
After this, the usual steps are agents and tool use, deeper retrieval work and guardrails. Basic Python helps, and the simulation lets you use a spreadsheet or Python on downloaded run, trace and section files.
The linked internships and free courses cover language model foundations, retrieval, prompt writing and AI security. The related roles are the machine learning engineer and the AI business analyst.
Keep the evaluation report, the incident note and the model and system card. Together they show you test what you build, which is the point job posts for this role return to again and again.
Keep going
Related job simulations and courses
Machine Learning Engineer
Train, evaluate and document a model, and catch target leakage.
AI Business Analyst
Use AI tools safely for real finance, HR and marketing work.
Data Analyst
Clean real-looking data, answer the business question and present it.
- Data and AI job simulations
- Free courses
- Browse jobs
- Virtual work experience in the UK
- Job simulation questions
Content reviewed 19 Oct 2026; UK context added 21 Oct 2026. Practice, not employment.
Questions
Quick answers
What does an LLM engineer actually do?
They build language model features into products: retrieval over company documents, prompts and tools, evaluation, safety checks and cost control. Roles differ by employer, so read each job description.
What is prompt injection?
It is when input to a model changes its behaviour in ways the developer did not intend. It is listed as the first risk in the OWASP Top 10 for LLM Applications.
Source: OWASP Top 10 for LLM Applications (opens in a new tab), checked 19 Oct 2026.
How do you evaluate an LLM assistant?
Write a test set of realistic questions with expected answers, score outputs against a rubric, and track failures when you change prompts or retrieval. Job posts ask for evaluation rubrics and prompt test datasets.
Do I need to code to do an LLM job simulation?
Basic Python helps, but every answer in the simulation can be found with a spreadsheet or with code on the downloaded files. The lab handles the retrieval tuning in the browser.
What is RAG?
Retrieval-augmented generation retrieves relevant passages from your documents and gives them to the model, so answers are grounded in the source. Employers ask for concrete evidence of it.
What should an LLM portfolio project include?
A working assistant, a test set with scores, a failure log and a security test. The recurring advice is to show evaluation, not only a demo.
How is this different from a prompt engineering course?
Courses teach techniques. A simulation gives you a brief, data and an incident to handle, then you document what you did.
What do LLM roles pay in the UK?
UK advertisements that mention LLM show a median salary of £85,000, with the middle half between £62,500 and £100,000, over the six months to 19 October 2026. This covers UK IT vacancies only. Pay in other countries varies, so no figure is given for them.
Source: ITJobsWatch, LLM (UK IT vacancies) (opens in a new tab), checked 19 Oct 2026.
Do I need to be able to code?
No. The tasks are checked as numbers, decisions and writing, and the hands-on exercise is done in EDWartens Lab. Python helps with the run and trace files, but a spreadsheet works for every calculation.
Is this a real company and a real assistant?
No. Halden Applied AI and Northfield Housing Association are made-up companies, the data is synthetic and the prices are invented. The documents, tickets and situations follow what a junior meets on a real project.
How long does it take?
The free week is about ten hours of work spread over five days. The full Premium edition runs over twenty working days and roughly thirty to forty hours, at your own pace.
Which model or vendor does it use?
None. The content is vendor-neutral. It talks about an LLM, a judge model and a vector index, so what you learn applies to any provider.
What do I get at the end?
A work record with your scores by task and skill, which shows on your EDWartens profile under Work experience, marked as a job simulation.
Is the AI and LLM Application Engineer job simulation free?
Yes, to start. Days 1 to 5 are free. The certificate (£14.99) and the Premium edition (£53.99) are optional, and both prices include VAT.
Does it count as work experience?
No. It is a training project at a simulated company, not employment, an internship or a placement. A job simulation is a realistic training project at a simulated company. It is not employment, an internship or a work placement, and it is not a job offer. Hours are active hours the learner logged on EDWartens UK.
A job simulation is a realistic training project at a simulated company. It is not employment, an internship or a work placement, and it is not a job offer. Hours are active hours the learner logged on EDWartens UK.
Start the AI and LLM Application Engineer job simulation today
Days 1 to 5 are free. No card to start. Practice, not employment.