EXEMPLUMeg-ZEM-plum
AI Portfolio Project
Exemplum is Latin for “a sample,” a piece of work held up to show what its maker can do. Ours is one AI project you build in four weeks, reviewed by an engineer, and yours to show.
Build an AI project in four weeks that a hiring team can open and check.
$1,500: four weeks, one project scoped to your target role, weekly code review, and a case study you can publish.
You write the code. An engineer reviews it each week and says plainly what a hiring team would question.
Is this for you?
- Career changers who need evidence that a hiring team can open and run
- Developers adding AI work to a portfolio of conventional projects
- Analysts and operators who want to show they can build with AI, not only use it
The problem
“Show me something you built” is the question that decides.
For AI roles, a hiring team wants to open something: a repository, a demo, a write-up with results.
Most tutorials end in a demo that cannot answer the next question: how well does it work, and how do you know?
One well-scoped project, measured honestly, answers it better than five unfinished ones.
How we fix it
Four weeks to a project you can show, with an engineer reviewing your code.
We scope one project for your target role, with a clear input, a clear output, and a way to measure it. It uses public or synthetic data, so you can publish it.
You build it. Each week an engineer reviews the code and the evaluation the way a colleague would, and sets the next week’s focus.
In the last week you write the case study and a demo walkthrough, including what does not work yet. Hiring teams trust that.
Projects by target role
Where a project fits.
Typical projects. Illustrative examples, not client results.
- AI engineer
An agent with an approval step
An agent that drafts replies from a document set, routes uncertain cases to a person, and reports its accuracy on a test set.
- ML engineer
A fine-tune against a baseline
A small model adapted to one task, compared with a general model on the same evaluation set.
- AI product manager
A specified and measured feature
A product spec, acceptance criteria, a working prototype, and a launch plan with quality gates.
- AI operations
A monitored workflow
An automated workflow with run logs, an exception queue, and a weekly quality report.
- Data analyst
An automated analysis
A recurring analysis rebuilt with AI assistance, checked against the hand-made version.
- AI governance
A risk review of a real tool
A written assessment of a public AI tool: data flows, risks, controls, and a sample policy.
What you receive
A scoped project
One project matched to your target role, with a clear input, a clear output, and a way to measure it.
Weekly code review
An engineer reviews your repository each week: structure, tests, evaluation, and the questions an interviewer would ask.
Evaluation built in
A small test set and a measure of quality, so you can say how well it works, not only that it runs.
A written case study
The problem, the design, the results, and what you would do next, edited for a hiring audience.
A demo walkthrough
A short script for showing the project in an interview or a recorded video.
How it works
Scope it
Agree the project, the data it uses, and how it will be measured. Public or synthetic data only.
Build, week one to three
You build. Each week an engineer reviews the code and sets the next week’s focus.
Measure
Run the evaluation, record the results honestly, including what does not work yet.
Write it up
Turn the project into a case study and a demo walkthrough you can show.
How the project is checked
What a hiring team will look for.
Each point can be checked in the repository or the case study.
- It runs
- Someone else can follow the README and run it.
- It is measured
- A test set and a quality measure, with the results recorded.
- It is honest
- The case study states what works, what does not, and what you would do next.
- It fits the role
- The project shows the skills your target role asks for.
Common concerns
What people ask before they buy.
- Data
- Public or synthetic data only, so the project is yours to publish.
- Your work, your code
- You write the code. Reviews explain and suggest; they do not do the work for you.
- Scope
- Four weeks suits one focused project. We cut scope before we cut evaluation.
Who does what
Your coach does the heavy lifting. You make the calls.
Your team
- Agree the project scope
- Build it, about eight hours a week
- Act on each week’s review
- Publish the result if you choose
Cursus
- Scope the project with you
- Review the code weekly
- Help design the evaluation
- Edit the case study and demo script
Afterward
Where people go next.
Most people take one of these routes. You choose, and nothing is booked for you.
Publish it
Put the repository and case study on your profile and résumé.
Practice presenting it
Mentor coaching can make it the subject of a scored mock interview.
Build the next one
Scope a second project for a different part of the role.
Before you start
What to have ready.
- Your target role
- The tools and languages you know
- About eight hours a week for four weeks
- A GitHub account
Related services
What the fee includes
Exactly what you get for the price.
- A project scope agreed in writing in the first week.
- Three weekly code reviews, written in your repository.
- An edited case study and a demo script in the final week.
- The code and write-up are yours to publish.
Plan my project
Tell us what you want to build.
We reply with a short scoping call and an invoice.
- A coach reads every request
- A reply by email within two business days of receipt
- Nothing is charged until you confirm
Not sure yet? Ask a coach first: a free 15-minute call about your move into AI.
Where this fits
- Start free Résumé grader
- Present Résumé & Profile
- Build AI Portfolio Project
- Grow Career Coaching
Not ready yet? AI career paths. After this: AI Résumé & Profile Rebuild or AI Career Coaching.
Questions
What kind of project?
Whatever your target role needs to see: an agent that drafts and routes work with a human approval step, a retrieval system with an evaluation set, or an analysis automated end to end.
Do I need to know how to code?
For engineering roles, yes. For product, operations, or analyst roles, the project can use no-code tools, and the review focuses on design and evaluation.
Can I use my employer’s data?
No. Projects use public or synthetic data, so you can publish them.
How is it paid for?
You request it here, and we reply by email with a scoping call and an invoice. Payment is taken before week one.