ai curriculum · training
the ai curriculum,course by course.
251 courses and 29 instructor-led workshops covering ai across engineering, product, data, sales, finance, legal, hr, operations, leadership and risk. 700.5 contact hours in total. every course lists its level, its hours, the tools it uses, and the artefact you walk out with.
this is not a video library. courses are taught in a browser environment with guided practice. labs are graded on what you produce. workshops are run live by the practitioners who ship our consultancy and platforms work.
courses across 16 topics
contact hours of taught content
instructor-led workshops, 65 sessions
graded labs. you build, we mark
Teams from across the business spend six weeks turning a real problem into a working AI tool, then present it to leadership.
A company wide build competition. Small teams pick a problem from their own work, build something that solves it over six weeks, and present it at the end to a panel who decide there and then whether to adopt it. Four live sessions guide the teams, with help available in between. This is the part of the programme where training turns into something the business can actually use.
the shape of it
- How long
- 6 weeks, 4 weeks, or a 48 hour hackathon
- Live sessions
- 4 sessions, 10 hours in total
- Support throughout
- A weekly drop in session every week, and a named mentor for every team
- Teams
- Up to 12 teams of 4 to 6 people
- People
- Up to 60, drawn from across the business
- Where
- Online, with the final presentations in person where possible
how long it runs
Six weeks
The recommended length. Four live sessions with two build periods in between, so teams have real time to act on the feedback from the first progress review. That correction is where most of the quality comes from.
Four weeks
The same four sessions run on a tighter calendar, with each build period shortened to a single week. It suits organisations that cannot hold people for six weeks, but teams have less time to change direction after feedback, so problems need to be scoped smaller at the kickoff.
48 hour hackathon
One continuous event across two days. Teams pick from the pre approved list only, because there is no time to define a problem from scratch. Mentors stay with the teams throughout rather than meeting them at review points, and the panel expects a working demonstration rather than a tested tool. This format is strongest at building momentum and surfacing ideas, and weakest at producing something you can put into service the following week.
what happens, week by week
the plan below is the six week version. the four week version runs the same four sessions, with each build period shortened to one week.
- Week 0Kickoff3 hour session
Teams form and choose what to work on. Each team can either take an idea from the pre approved list drawn up with sponsors before the challenge starts, or put forward an idea of its own. Ideas submitted by a team are reviewed on the day, and once approved the team is free to take either route. Whichever they pick, the problem is then cut down to something they can genuinely finish in the time available.
by the end of this stageAn approved problem, taken from the list or submitted by the team, and a clear description of what a good result would look like.
- Weeks 1 to 2Teams buildWeekly drop in session, plus a mentor
Teams build their first version. Stennir runs a drop in session each week, and every team has a named mentor they can reach in between.
by the end of this stageA first version that runs, however rough it looks.
- Week 2First progress review2 hour session
Each team shows what they have built so far and gets feedback. This is where the scope gets corrected, and some teams are advised to drop half of what they planned.
by the end of this stageA shorter and more realistic plan for the weeks that remain.
- Weeks 3 to 4Teams buildWeekly drop in session, plus a mentor
Teams keep building, with the weekly sessions now focused on testing and on what the tool does when it gets something wrong.
by the end of this stageA set of test cases, and the first safety checks in place.
- Week 4Testing and safety review2 hour session
Teams measure how well the tool works, what it costs to run, and how it handles unclear or incorrect input. A Stennir reviewer works through each tool with its team and raises the questions the panel is likely to ask.
by the end of this stageA tested tool, answers to the review questions, and a plan for how it would be used day to day.
- Weeks 5 to 6Final preparationRehearsal with your mentor, plus drop in support
Teams finish building and gather the evidence that their tool works. Every team rehearses its presentation with its mentor and gets feedback on how to make the case clearly in five minutes.
by the end of this stageA rehearsed five minute demonstration and a one page summary.
- Week 6Final presentations3 hour session
Each team demonstrates their tool live to a panel of sponsors and senior leaders, then answers questions. The panel scores against the criteria below.
by the end of this stageA decision from the panel: adopt the tool, keep working on it, or stop.
the 48 hour hackathon schedule
The 48 hour format keeps the same judging criteria, with one change: the weight on evidence moves towards the demonstration itself, because two days is not enough to build a proper set of test cases. Teams whose idea is worth pursuing usually then run the six week version to turn the prototype into something usable.
- Day 1, 09:00Kickoff1 hr
Briefing on the rules and the judging criteria, teams form, and each team picks an idea from the pre approved list.
- Day 1, 10:00Build7 hrs
Teams build. Mentors circulate throughout rather than waiting for a scheduled review.
- Day 1, 17:00Checkpoint1 hr
Each team shows what actually runs. Scope gets cut here, which is the difference between finishing and not.
- Day 1, 18:00Optional evening sessionoptional
The room stays open for teams that want it. Nobody is expected to stay.
- Day 2, 09:00Build5 hrs
Teams finish the working parts and drop anything that will not be ready.
- Day 2, 14:00Safety check1 hr
A mentor works through each tool with its team, covering data handling and the obvious ways it could go wrong.
- Day 2, 15:00Final build and rehearsal2 hrs
Teams stabilise the demonstration and prepare a short case for why it matters.
- Day 2, 17:00Demonstrations and judging2 hrs
Each team demonstrates live to the panel, which scores and decides which ideas are worth taking further.
how the panel scores each team
- 30%Business impactIs the problem real, and would solving it matter to people outside the team?
- 25%A working toolDoes it run on real data, live in the room, rather than as a rehearsed demonstration?
- 20%Evidence it worksAre there test cases, a measured result, and an honest account of where it gets things wrong?
- 15%Safe and responsible useHow data is handled, what the tool refuses to do, and whether users can tell AI is involved.
- 10%Plan for using itWho owns it next, what it costs to run, and what needs to be true before it can be used properly.
who is in a team
Teams are deliberately cross functional. A team that is all engineers builds the wrong thing well, and a team with no engineer does not build anything. Each team should bring together the perspectives below, in whatever mix suits the problem it has chosen.
- Someone technicalBuilds the tool and makes the technical decisions
- Someone who owns the problemDefines the problem clearly and can show what solving it is worth
- People who do the work day to dayProvide the practical knowledge and judge whether the answers are actually right
- A manager as sponsorApproves the problem, attends the final presentations and decides what happens next
what you get out of it
outcomes
- Up to twelve working tools, built on the organisation's own data and systems
- A ranked list of those tools, with a decision recorded against each one
- People who have now built something with AI rather than only learned about it
- A written record of real, tested uses that later projects can build on
- A clear picture of which teams can build on their own and which need more support
before you start
- Everyone taking part has finished the AI Foundations courses
- Each team includes at least one person at Advanced level
- A list of pre approved ideas has been agreed with sponsors before the kickoff
- Each team has a manager who sponsors the problem and attends the final presentations
- Problems come from work that is already causing difficulty, not from a brainstorm
who provides what
stennir provides
- Design of the challenge, facilitation on the day, and the scoring criteria
- Help drawing up the pre approved list of ideas with your sponsors beforehand
- Four live sessions plus a weekly drop in help session across the six weeks
- A named mentor for every team
- Practice environments set up with data similar to your own
- A written report at the end covering the decisions taken and how the teams performed
you provide
- A shortlist of pre approved ideas, agreed with sponsors before the challenge starts
- Managers willing to sponsor those problems and to review any new ideas teams put forward
- Time for participants, roughly four hours per person per week
- Access to the data and systems the teams need, agreed before the kickoff
- A panel for the final presentations, including someone who can approve adoption
want a track built for your team?
tell us the team, the outcome and the calendar. we come back with a track drawn from this catalogue, the cohort dates, and the price.