FREE PLAYBOOK · PAWANAX CAREERS
From laid-off to lead teacher: a playbook for turning years of craft into AI-era teaching
A lead teacher, in this playbook, is an experienced practitioner who designs a lesson around real work, demonstrates it done well and done badly, and checks whether learners can do it without help. Lead teachers keep the final say on what good looks like in their domain.
This playbook explains how to build that lesson from years of craft, how classic roles map to AI-era work, how to keep former-employer material out of your teaching, and how to show AI literacy in one weekend. It is free to read and assumes no teaching or coding experience.
1. Turn your craft into a 90-minute lesson
A good lesson starts from one outcome a learner can demonstrate, not from a list of tools.
Professionals who try to teach often start with the tool. Start with the work instead. Write one sentence that begins "By the end, the learner can..." and make it observable: a checked claim file, a reconciled ledger, a tested prompt set, a reviewed drawing.
Use the template below for every lesson. The time boxes add up to 90 minutes.
The 90-minute lesson template
| Part | Minutes | What you do |
|---|---|---|
| Outcome | 5 | State the observable result in one sentence and name the real artefact the learner will produce. |
| Demo 1: the normal case | 15 | Do the task the way you do it at work, narrating each decision aloud. |
| Demo 2: the variation | 15 | Change one input (a different customer, document or language) and show what changes. |
| Demo 3: the AI-assisted version | 15 | Run an AI tool on the same task. Show the prompt, the output and the three checks you run before you trust it. |
| Failure case | 10 | Show one realistic mistake, where it came from, and the check that would have caught it earlier. |
| Exercise | 20 | Learners do one small version of the task on practice material, alone or in pairs. |
| Review and close | 10 | Learners name what they would check first next time, and you collect one open question. |
Three rules keep the demos useful:
- Use practice material only. Never demonstrate on live client, patient or employer data.
- Narrate judgement, not clicks. Say why you chose the option you chose.
- One idea per demo. If a demo needs a second idea, it is two demos.
The failure case is the most valuable block. It shows learners that AI output can be plausible and wrong at the same time, which is the skill a seasoned practitioner already has.
2. Role translation: from classic roles to AI-era work
Much AI-era work is a new name for a judgement you already use: deciding whether an output is right, and what to do when it is not.
Job titles differ between employers and change quickly. The table shows the kind of work each classic role maps to and the judgement that carries over. Treat the middle column as examples of work, not as a guaranteed opening.
Twelve classic roles and the AI-era work that uses the same judgement
| Classic role | AI-era work that uses the same judgement | Judgement that carries over |
|---|---|---|
| Claims adjuster | AI output reviewer for claims triage | Knowing which details decide the outcome, and when a reasonable-looking decision is wrong |
| Paralegal | Document-review evaluator for legal workflows | Chronology, citation discipline and spotting the clause that is missing |
| Radiologic technologist | Imaging data-quality reviewer, working under clinical supervision | Quality standards and telling an acquisition artefact from a real finding; the clinician keeps the diagnosis |
| Logistics planner | Operations workflow designer for exception handling | Constraint thinking and knowing exactly where a plan breaks |
| Accountant | Reconciliation reviewer for AI-prepared books | Reconciliation, audit trails and deciding what is material |
| QA engineer | Domain red-teamer and evaluation designer | Designing test cases, adversarial thinking, and grading severity |
| Recruiter | Rubric author and reviewer for structured screening tools | Consistent criteria, bias checks and documenting why a decision was made |
| Teacher | AI curriculum designer and faculty lead | Sequencing, useful feedback and checking that understanding has actually happened |
| Journalist | Source-verification reviewer for AI-drafted text | Sourcing, verification habits and a disciplined correction process |
| Translator | Localisation and multilingual output reviewer | Meaning over literal wording, register, and keeping a glossary consistent |
| Customer-support lead | Conversational-AI quality lead and escalation designer | De-escalation, clear escalation paths and a tone policy people can follow |
| Mechanical drafter | Technical documentation reviewer checking outputs against specifications | Tolerances, standards, and reading a specification against its intent |
To use the table, find your row and write your own version of the middle column in one line. Then name the one judgement you would bring into your first week.
3. Clean room: bring your craft, not your former employer's secrets
You can teach everything you learned through practice without carrying a single confidential document.
A clean room is a set of rules for building teaching material from what you know, not from what you kept. Your method, your general judgement and your checklists, written fresh, are yours. Internal files, customer records, pricing sheets, source code, unreleased plans and screenshots of internal systems are not.
- Do not copy, photograph or forward internal documents, customer data, pricing sheets, source code, dashboards or internal chat logs.
- Rewrite every example from memory as a generic case. Remove names, numbers, identifiers and anything that points to one employer.
- Use public or synthetic data for every demo and exercise, and label synthetic data as synthetic.
- Read your former employer's agreements, including confidentiality, non-solicitation and return-of-property terms. If a term is unclear, ask a qualified lawyer before you publish anything.
- Describe methods in your own words, as how you work, not as a copy of an internal procedure.
- Keep a separate folder for teaching material, with no file that came from a former job.
- Ask a peer to read your lesson for anything that looks internal before it is shared.
- Your employment situation is not a required part of any lesson or profile. Share it only if you choose to.
This checklist is general practice, not legal advice. Pawanax faculty agree to clean-room terms in writing before any work starts.
4. Prove AI literacy in a weekend
AI literacy is shown by artefacts you can point to: a verification log, an evaluation set, a documented workflow, a failure analysis and a taught walkthrough.
Five tasks, each scoped to a few hours, make a folder of dated artefacts that anyone can review. No coding is required. Use generic or synthetic material throughout, so the folder stays clean-room compliant.
- Verification log. Take one real task from your past craft, rewritten generically. Ask an AI tool for a draft. Find five errors or unsupported claims, and for each one record how you found it and which check would have caught it earlier.
- Small evaluation set. Write 20 cases from your domain, each with the answer you would accept and one reason it is right. Run the same cases through two prompts or two tools, score both, count the disagreements and explain three of them.
- One documented workflow. Automate one repetitive step with a spreadsheet formula, a no-code tool or a short script. Document the inputs, the outputs, the human check, and the two most likely failure modes.
- Failure analysis. Find one confident wrong answer and classify its cause: missing context, ambiguous instruction, a bad source, or a case the tool was never built for. Write the fix you would try first.
- Ten-minute teach-back. Record a five-minute walkthrough of tasks 1 to 4 for a peer. Write down the one question you could not answer and what you would check next.
Done means a folder with five dated artefacts. Share it with a peer or a lead teacher, and ask them to point at the weakest one.
5. How Pawanax faculty works
Faculty teach real practice in a voluntary launch-phase role, under written terms that arrive before any work starts.
Pawanax Academy recruits experienced professionals as human faculty to teach practical AI. Learners are guided by AI coaches, and the work is signed off by real people.
Voluntary launch-phase role. Pawanax is not paying faculty or interns at this time, and we say so up front.
What the role offers, in writing:
- Public credit and a named profile.
- A portfolio and teaching evidence you can show later.
- Structured upskilling into AI roles, with each step written down.
- References on request.
- Flexible hours, planned around your existing work.
- Clean-room protection: you never have to bring a former employer's confidential material.
Written terms arrive before you commit. They cover scope, start date and cancellation. No work starts before those terms are agreed in writing, and Pawanax does not promise jobs, renewals or hours.
Read the faculty roles and their written terms.
6. Frequently asked questions
Short, direct answers to the questions candidates ask most.
What is a lead teacher?
A lead teacher is a practitioner who designs a lesson around real work, shows it done well and done badly, and checks whether learners can do it without help. Lead teachers keep the final say on what good looks like in their domain.
Do I need to code to teach AI-era work?
No. The AI-era work in this playbook is evaluation, workflow design and review, which use documents, spreadsheets and judgement. Scripts are optional, and task 3 of the weekend checklist accepts a no-code tool.
Do I have to mention a layoff?
No. Your employment situation is optional and private. Lessons and profiles are judged on the work, and sharing anything about your situation is your choice.
Can I use what I learned at my last job?
You can use the method and the judgement, but not the documents. Follow the clean-room checklist in section 3, and ask a qualified lawyer about any agreement you are unsure of.
How long does the weekend take?
Plan one weekend. Each of the five tasks is scoped to a few hours and uses generic or synthetic material, so you can stop after any task and resume later.
How do I apply, and what do I need?
Start at the faculty roles page. An application needs your name, your email, a PDF CV of up to 10 MB, a short note on why the work matters to you, and your explicit consent. No account is required, and there is no fee to apply.