Training
Agentic Coding Training for Engineering Teams
Hands-on, practitioner-led training that gets your developers productive with agentic coding tools — Claude Code, OpenAI Codex, and the wider ecosystem of AI-native developer tooling. Not a lecture, not toy examples: your team ships real work on your real codebase, and leaves with the workflows, guardrails, and habits that turn these tools into a durable productivity advantage.
- Your codebase, not toy examples
- Every exercise runs against your team's actual repositories and stack.
- One flat fee per cohort
- No per-seat pricing. Train up to 12 developers for a single fixed price.
- Remote or on-site
- Delivered live over video worldwide, or in person across the UK and US.
What your team will be able to do
By the end of the workshop, every participant has shipped meaningful work with an agent — and knows how to keep doing it after we leave.
- Drive agents through real features.
- Plan, implement, and verify multi-file changes in a production codebase — not just autocomplete snippets.
- Engineer context deliberately.
- Write project instructions (CLAUDE.md, AGENTS.md) that make agents effective in large, unfamiliar codebases.
- Run test-driven agent loops.
- Use tests, linting, and CI as guardrails so agents verify their own work instead of you babysitting them.
- Extend agents with MCP.
- Connect agents to your internal tools, databases, and services with Model Context Protocol servers.
- Review agent output at scale.
- Keep code review, standards, and quality intact when the volume of generated code goes up.
- Adopt safely in regulated environments.
- Apply permission controls, secrets hygiene, and governance patterns that satisfy security and compliance teams.
The curriculum
A full day, hour by hour
We publish exactly what we teach — no vague module titles. Before the workshop we review your stack and repositories, so every session below runs on your code. Timings flex to fit your team's day.
-
09:30 — 10:30
Module 1 · Foundations & setup
- The agentic tooling landscape: Claude Code, OpenAI Codex, and IDE-based agents — where each shines and how to choose
- How agents differ from autocomplete: the mental-model shift that makes or breaks adoption
- Installation, authentication, and permission modes on every participant's machine
- First exercise: a real task, end-to-end, in your codebase
-
10:30 — 11:45
Module 2 · Context engineering
- Writing project instructions that work: CLAUDE.md, AGENTS.md, and equivalents
- Structuring prompts and context for large, unfamiliar codebases
- Planning before implementing: when to explore, when to plan, when to act
- Exercise: author a project instructions file for one of your repositories, then measure the difference it makes
-
12:00 — 13:00
Module 3 · Core workflows
- Test-driven agent loops: letting tests, linters, and type checkers verify the agent's work
- Refactoring and working in legacy code with agents
- Debugging with agents: logs, reproduction, and root-cause analysis
- Git workflows: commits, branches, and pull requests driven by agents — with humans in control
-
14:00 — 15:15
Module 4 · Advanced workflows
- MCP servers and custom tools: connecting agents to your internal systems
- Subagents and parallel sessions: orchestrating multiple streams of work safely
- Hooks, automation, and slash commands: encoding your team's standards into the tools
- Agent-assisted code review and CI integration
-
15:30 — 16:30
Module 5 · Guardrails & governance
- Permission models, sandboxing, and blast-radius control
- Secrets hygiene and keeping sensitive data out of agent context
- Reviewing agent output at scale without rubber-stamping
- Adoption patterns for regulated industries: finance, healthcare, pharma
-
16:30 — 17:00
Wrap-up · Your 30-day adoption plan
- Team norms: what good agentic engineering looks like for your organisation
- Measuring impact: metrics that show real productivity gains, not vanity numbers
- A concrete rollout roadmap for the next sprint and the next month
Everything your team leaves with
- A written agentic coding playbook tailored to your stack
- Project instructions files written during the day, ready to commit
- A governance & permissions checklist for security sign-off
- A 30-day adoption roadmap with metrics to track
- A full recording of the session for future joiners
- Two weeks of follow-up questions answered by email
Formats & pricing
Simple, transparent, per cohort
Most training in this space is priced per seat or hidden behind a sales call. Ours isn't: one flat fee per cohort, published here. Every format is delivered live by practitioners who ship production software with these tools every day.
Leadership briefing
Half day · for CTOs, VPs, and engineering managers
$1,950 / session
Approx. £1,500 · delivered live & remote
- The agentic tooling landscape, demystified
- ROI, risk, and governance for AI-assisted engineering
- Live demonstrations on real production code
- A decision-ready adoption strategy summary
- Up to 15 leaders included
Most popular
Team workshop
Full day · the complete curriculum above
$4,500 / cohort
Approx. £3,500 · flat fee, delivered live & remote
- Up to 12 developers included — under £300 per head
- Full day of live, hands-on instruction
- Tailored to your codebase & stack before the day
- Playbook, recording & all deliverables listed above
- Two weeks of follow-up support included
Additional participants $250 each. On-site delivery available — travel billed at cost.
Book this workshopEnablement programme
Workshop + 4 weeks of embedded support
$9,500 / team
Approx. £7,400 · multi-team programmes quoted on request
- Everything in the team workshop
- Four weekly live office-hours sessions after the workshop
- Async support in your Slack or Teams throughout
- Playbook & tooling config refined with your team in the wild
- End-of-programme adoption & impact review with leadership
How we compare
There's no shortage of AI training. Most of it is either a generic classroom course taught on toy examples, or a video library nobody finishes. We built this workshop to be the one that actually changes how your team works on Monday morning.
| Drutek | Generic classroom course | Self-paced video course | |
|---|---|---|---|
| Taught on your own codebase | Toy examples | Toy examples | |
| Led by practitioners who ship production software with these tools | Usually professional trainers | Varies | |
| Flat per-cohort pricing, published up front | Typically per seat | Per seat | |
| Governance & compliance module for regulated industries | Rarely | Rarely | |
| Deliverables you keep: playbook, instructions files, checklists | Slides | Videos | |
| Follow-up support after the session | Rarely | — | |
| Updated as the tools change, week to week | Fixed syllabus | Ages quickly |
Who it's for
Built for engineering organisations that want to adopt agentic coding deliberately — with real productivity gains and without cutting corners on security or quality.
Software teams
Developers who want to move faster with agents while keeping code review, testing, and standards intact. Any language, any stack — the workflows transfer.
Engineering leaders
Managers rolling out AI tooling who need consistent, safe adoption across teams — and a clear picture of the ROI to show the board.
Regulated organisations
Teams in finance, healthcare, and pharma that need agentic workflows to fit within existing governance controls — a core module of the curriculum, not an afterthought.
Frequently asked questions
- Do we need tool licences before the workshop?
- We'll tell you exactly what to have in place during a short pre-workshop call — typically subscriptions or API access for the tools your organisation has chosen. If you haven't chosen yet, we'll help you pick, and we can run the day across more than one tool.
- Our code is sensitive. How does training on our codebase work?
- We're happy to sign an NDA, and all exercises run on your machines and in your environment — your code never passes through ours. For highly restricted environments we can run the day on a representative internal repository or a sanitised mirror instead.
- Which tools do you cover?
- We're tool-agnostic. The day is usually anchored on Claude Code and OpenAI Codex, with IDE-based agents like Cursor and GitHub Copilot covered where teams use them. The workflows and habits we teach transfer across all of them.
- What experience level is required?
- Participants should be working developers comfortable in your codebase. Mixed-seniority cohorts work well — juniors and principals get different things from the same exercises, and we calibrate on the day.
- What if our team is bigger than 12?
- Additional participants are $250 each, though we find cohorts of 8–12 get the most hands-on attention. For larger organisations we run multiple cohorts or a tailored multi-team programme — talk to us and we'll quote it.
- Remote or in person?
- The published prices are for live remote delivery, which works well for hands-on work. We also deliver on-site across the UK and US — same fee, with travel billed at cost.
Ready to upskill your team?
Book the Agentic Coding workshop, or talk to us about a leadership briefing or a tailored multi-team programme.