SQValex Guides
Free, practical guides on AI engineering, quality engineering and AI-enabled delivery, by course, from beginner to advanced, written by the academy that teaches them. Every guide is free to read and written from the lessons our learners work through.
12 guides · 6 courses · 2 tracks
New to all of this? These 4 guides make sense without any background, and each one stands on its own.
Generative & Agentic AI · How AI works
A plain-language walkthrough of how LLMs work, with three worked examples you can check by hand: tokens and cost, attention, and temperature and top-p sampling.
Beginner18 min readUpdated 6 November 2026
Forward Deployed Engineer · AI roles and careers
A forward deployed engineer makes AI work inside a client's real business. Here is what the role involves, how it differs from ML engineers and architects, and its skills.
Beginner5 min readUpdated 6 November 2026
Delivery Foundations · Using AI safely
Six checks to run before an AI-written draft goes into delivery work: claims, numbers, citations, personal data, hidden instructions and who signs it off.
Beginner14 min readUpdated 7 November 2026
Business Analysis · Requirements and analysis
AI can prepare interviews, summarise transcripts and flag gaps, but it cannot decide what is true. Learn what to delegate, what to keep and a workflow that holds up.
Beginner12 min readUpdated 7 November 2026
You build it, measure it, harden it and ship it. Most graded work happens inside a real or simulated AI system.
A 16-week course on one realistic application: test design, API and security testing, automation and CI, and testing AI itself. Every practical is checked by code.
Test design
Learn equivalence partitioning, boundary value analysis, decision tables, state transitions and pairwise testing, with worked examples and how to write a good test case.
Beginner17 min readUpdated 6 November 2026
AI and testing
How testers use AI well without trusting it blindly, and how to test machine-learning models, LLMs, RAG and agents: precision, recall, metamorphic tests and OWASP risks.
Intermediate17 min readUpdated 6 November 2026
A 12-week course on one enterprise engagement: scope it, deploy it privately, build retrieval and agents, guard it, evaluate it and hand it over. Every lab is checked by code.
AI roles and careers
A forward deployed engineer makes AI work inside a client's real business. Here is what the role involves, how it differs from ML engineers and architects, and its skills.
Beginner5 min readUpdated 6 November 2026
Retrieval and RAG
How production RAG systems combine keyword search and embeddings, fuse the results with reciprocal rank fusion, rerank a shortlist and measure the result with recall and MRR.
Intermediate5 min readUpdated 6 November 2026
An 8-week, self-paced course with 25 auto-graded labs and a short video each week, from how a model works to launching an assistant in production.
How AI works
A plain-language walkthrough of how LLMs work, with three worked examples you can check by hand: tokens and cost, attention, and temperature and top-p sampling.
Beginner18 min readUpdated 6 November 2026
Retrieval and RAG
How retrieval-augmented generation works end to end: chunking, BM25 and embeddings, hybrid fusion, reranking, citations, abstention and how to evaluate it.
Intermediate16 min readUpdated 6 November 2026
Use AI to go faster without losing control: verify what it drafts, keep personal data out of it, set where it may act, and own the decision.
A 4-week course for anyone on a software delivery team: check an AI draft claim by claim, keep personal data out of it, ask properly and move a change through six delivery gates. Twelve labs are graded by the platform.
Using AI safely
Six checks to run before an AI-written draft goes into delivery work: claims, numbers, citations, personal data, hidden instructions and who signs it off.
Beginner14 min readUpdated 7 November 2026
Prompting and context
How to ask an AI assistant for project and delivery documents: write a brief, supply approved sources, fix the format, say what to do when information is missing, then verify.
Beginner12 min readUpdated 7 November 2026
An 8-week course in which AI drafts and you own the truth: discovery, processes, stories, rules, data, non-functional requirements and change impact, ending in a sign-off pack of requirements you can test and trace. Twenty-four labs are graded by the platform.
Requirements and analysis
AI can prepare interviews, summarise transcripts and flag gaps, but it cannot decide what is true. Learn what to delegate, what to keep and a workflow that holds up.
Beginner12 min readUpdated 7 November 2026
Requirements and analysis
What makes a story and its acceptance criteria good, how to prompt AI for drafts, and a checklist for catching invented rules, missing cases and vague words.
Intermediate13 min readUpdated 7 November 2026
An 8-week course on deciding with evidence while AI drafts and calculates: outcomes, briefs, discovery, backlogs, prioritisation, AI features, experiments and release scope. Twenty-four labs are graded by the platform.
Product ownership
How a product owner separates outcomes from output and AI-adoption counts, writes a brief from evidence, tests assumptions and says how a customer is better off.
Beginner13 min readUpdated 7 November 2026
Product ownership
What an AI assistant can do for your backlog and what you must decide yourself, with RICE, cost of delay and WSJF worked through and the order rechecked.
Intermediate14 min readUpdated 7 November 2026