你的职业操作系统Your Career Operating System
Don't just learn. Build your career.
CareerOS runs on your own computer and does real work through your own Claude Code or Codex. It starts by reading your resume and the role you want, and goes all the way to delivered work and application materials. Each step leaves versioned evidence an employer can check.
macOS release · Apple silicon and Intel · signed and notarized
“Built a task tracker REST API with Node.js, Express and PostgreSQL, with Jest tests.”Resume · Riley_Sample_Resume.pdf
Task 2 · Fix API error handling
Tests: 14 passed, 14 totalThe AI check only describes what it observed in this version
Stuck? Ask me below and I'll start with the smallest hint.
CareerOS checks the job ad line by line against what you have, lists what's there and what's missing, and gives the source for each point.
An accounting master's student wants a Junior Data Analyst role. The resume has no SQL; the ad asks for strong SQL.
In the CareerOS desktop app you work on tasks in a real project, and your code changes and test results become work an employer can check.
Task 2 · Fix the API error handling · Tests: 14 passed, 14 total
First check how much AI affects the job you do now, then look at the roles you could move into and what each one needs.
Both are free to read: the AI impact on existing careers, and the new AI-era roles with steps for moving into them.
The order never changes: Assess, Plan, Act, Prove, Connect, Grow. After Grow you take the new feedback back to Assess.
Step one, Assess, and step two, Plan, are free on the JR Academy website. From step three, Act, Prove, Connect and Grow happen in the CareerOS desktop app.
What you doDescribe where you are and the role you want. Add your resume or a job ad and the picture gets sharper.
What you end up withA list of where you stand and what's missing, each point with its source.
What you doOrder the gaps and fit them into the hours you actually have each week, with the matching course, certification or project.
What you end up withA week-by-week plan that says what each week produces.
What you doWork on tasks in a real project, pin your code changes, test results and write-up as versions, match them against a job ad for your application, and log interview feedback.
What you end up withA portfolio employers can check, and application material to go with it.
This is a demo of the Assess stage: you drop material into the chat, CareerOS reads it on your machine, and the AI turns where you stand into a card that quotes the source for every line. When you confirm, the steps on the right tick themselves off. The person and resume are fictional.
Illustrative animation, not a screen recording · fictional person and resume
The top of the app always shows your career profile and these six stages. Open a stage and the centre becomes that stage's conversation, while the right panel lists the steps for it. Grow records interview results and feedback, then takes you back to Assess or Plan with new information.
What you do here
Drop your resume (PDF or DOCX) and a job description into the chat. The AI summarises where you are, where you want to go and the gaps between, quoting the source for every item. You confirm each one.
What it leaves behind
A sourced picture of where you stand: strengths, unknowns, and constraints such as time and location.
Steps on the right
The first complete track is Junior Software Engineer: 12 tasks over about 5 weeks. These people are the best fit to start now.
You have learned a back-end language and Git and built at least one course or personal project, and want to turn it into work an employer can open and check.
You can set a career change as your goal. CareerOS reads your past experience and the role you want, finds the gaps, then plans the tasks.
The tutor guides you in Chinese, while resume bullets and interview stories are prepared for English-language roles, such as a 180–230 word STAR interview script.
All three are useful. This table compares how they work, not which is better. CareerOS also maps your gaps to JR Academy courses, and runs tasks through your own Claude Code / Codex.
| Compared on | Taking a course | Chatting with ChatGPT / Claude | CareerOS |
|---|---|---|---|
| Starts from your target role | Follows the syllabus; everyone in the cohort gets the same content | Possible, depending on how you describe yourself and the role each time | Starts from your resume and the role you want, finds gaps, then plans |
| Work in a real project | Assignments and projects scoped by the course | Discuss and generate code in a chat window, then move it into your project yourself | Local Claude Code / Codex reads code and runs tests in a project you authorise |
| Leaves checkable evidence | Grades and a completion certificate | Chat history you organise yourself | Diffs, test results and your explanation pinned as versions in a portfolio |
| Guidance at every step | Teachers and tutors on a timetable | Answers as far as you ask | Each step states its done criteria; the tutor reminds you what comes next |
| Keeps looping | Runs for the length of one course | You keep track of goals and progress yourself | Grow records interview feedback, then returns you to Assess or Plan |
Every stage in CareerOS is work you do with an agent. The centre is the conversation, where everything gets done; the right panel holds that stage's steps and walks you through them.
Resumes can be PDF or DOCX; a job description can be a link, a file or pasted text. The AI reads them and replies with a structured proposal instead of asking you to fill in a long form.
Proposals appear as cards in the chat. You confirm, edit or reject each item, and each one quotes the text it came from. Nothing is saved to your profile until you confirm. Work rights are entered by you only; the AI never reads or guesses them.
Each step on the right states its done criteria and why it matters. The AI tutor reminds you what to do now and, when you are stuck, starts with the smallest hint rather than writing it for you.
"Familiar with Node.js" on a resume is hard to verify. CareerOS records what you actually did in a project you authorised, and keeps "activity observed" separate from "skill confirmed".
Code diffs, test results, commits and the output of commands CareerOS ran for you, plus the explanation you write yourself.
Each pin creates a new version and keeps the old ones. Your portfolio points to a specific version, so later code changes never quietly alter what someone has already seen.
The AI check only states what it observed in that version and never raises the evidence tier. Human review is a later-stage service; reviewers and arrangements are not yet decided.
CareerOS packages abilities such as resume analysis, JD matching and cover letters as skills, then picks the most specific version for your role family × level × target market. If no specific version exists it falls back to the general one and says so in the chat.
Software Engineer at every level × Australia comes first. Other combinations use the general version, labelled as such under each result.
Criteria checked one by one
The AI's first suggested fix
“This is the only evidence on the whole resume, but it only lists technologies. It doesn't show what problem it solved, what you were responsible for or the hardest problem, and a reviewer can't open the project.”
Each criterion is only judged as met, partly met or not seen. CareerOS counts them; it is not a score.
The same fictional resume is analysed under three career profiles. The suggestions shown are translated from the AI’s original Chinese responses.
These are entry points you can see in CareerOS. Accounts, purchases and usage stay in each product; nothing has been merged.
Recommended in Plan for your skill gaps, taken during Act.
Certificates a target role asks for appear in your plan; one you've passed is proof an employer can check.
Build teamwork experience, recorded as a team project, not an internship.
Pick roles, write resumes and applications against the requirements, and prepare interviews with a mentor.
Meet people heading the same way; hackathons also add project experience.
30 days is a series. Each course is a planned 30 days: one theme, Day 1 to Day 30, a small deliverable every day, moving forward by date.
Courses in preparation are not on sale yet and have no price.
CareerOS for macOS is available to download now, for Apple silicon and Intel Macs; the first pilot is still being prepared. Here is what you can do today, and what comes later.
Pilot places, dates and pricing have not been announced. Use the email, phone or consultation options on our contact page to tell us the role you are aiming for, and we will reach out when the pilot opens.
Assessment and planning are free on the JR Academy website: describe where you are and the role you want, and you get a list of gaps with their sources and a week-by-week plan. Act, Prove, Connect and Grow happen in the CareerOS desktop app: you work on tasks in a project you authorise, pin code changes, test results and your write-up as evidence, then write applications against a job ad and log interview feedback.
Not for assessment and planning on the website. The first complete route in the desktop app is Junior Software Engineer, and its tasks happen in a code project, so you need to be able to write code already. Other role families use the general skill versions for now, and the result says so.
A Mac, plus either Claude Code or Codex signed in with your own Claude or ChatGPT paid plan or API credits. Model costs for running tasks are billed to your account and never pass through CareerOS; the AI tutor is provided by CareerOS.
Only macOS for now, for Macs with Apple Silicon (M-series chips). The installer is signed and notarised by Apple. Windows and Linux versions don't have a release date yet.
Your career profile, resume text and evidence records stay on your own computer. The agent can only read project folders you authorise, and read-only tasks never get write access. An exported portfolio PDF lands in your Downloads folder, and you decide who receives it.
CareerOS subscription pricing and the free tier have not been decided; they will be announced during the pilot. 30 days courses are bought separately: the first, Master AI in 30 days, is A$69 + GST as a one-off purchase.
No, and we make no job or income promises. CareerOS turns your learning, projects and job-search actions into a process backed by evidence; interviews and offers depend on the market, the role and you.
The first complete track is Junior Software Engineer: 12 tasks over about 5 weeks. Role families such as Data Analyst, Business Analyst, DevOps / Cloud, AI Engineer and Product Manager start with general skills, and results state that they are not yet calibrated for your role and level.
Courses fill capability gaps. Confirmed gaps in your career profile point you to matching JR Academy courses, 30 days lessons or hands-on labs, and you can open your enrolled courses from the CareerOS sidebar.