Twelve months, mapped so the biggest gaps close first.
This is not a list of good ideas. Each month is tied to a specific gap on your evidence page, sequenced so the earliest work makes the later work easier, with the breaks between terms doing real jobs of their own. Still in classes or already graduated, the sequence works the same.
The three blocks below map to your remaining terms, and the breaks are real breaks. Course picks like Research Methods slot into registration as it comes.
Same sequence, no syllabus required. Month one starts the day you decide, an online course or workshop stands in for the class pick, and the breaks become catch-up weeks.
Fall term
Sep to Dec 2026NowRest is allowed. The one thing that carries over: a cleaned-up portfolio link, ready for the January applications.
Spring term
Jan to Apr 2027BuildThe second UX research internship, May to Aug. Twelve weeks of real sessions and real users, producing the numbers your senior-year applications lean on.
Fall term, senior year
Sep to Dec 2027RecruitingThe order changes when your evidence does
Answer a prompt and a step can drop off. Change your target career and the whole sequence rebuilds around the new durable core. This is the reason the product is a subscription rather than a report: a plan written once for a junior is wrong by the following autumn.
What is an evidence gap?
An evidence gap is part of your target role's durable work that nothing on your record backs up yet. Coverage is the other side of it: how much of that work you can already show with something you actually did, out of 100. The durable work is the part the AI-exposure score says current AI cannot do well, built as described on the methodology page.
The sample has three kinds of gap, and each gets a different fix. Framing is a skill gap, so it gets a course. The onboarding case study backs a resume claim that had nothing behind it, so it gets a piece of work. Presenting is an open prompt, so it gets a talk at the showcase. Each one closes when the evidence page records it, which is why the March card says to note the audience size.
How is each month's work chosen?
Each month goes to the largest open gap on your evidence page that the calendar allows right now. In the sample, framing is the largest gap, so Research Methods starts in September and runs the whole term; the case studies, the internship, and the interviews later in the year all lean on it.
Work with an outside deadline goes in the month that deadline sets: internship applications in January, the showcase talk in March. The breaks do their own jobs. Winter break carries one task, a cleaned-up portfolio link, and summer is the internship itself.
What happens to a step that is still open at the end of a term?
It moves into the next term. In the sample, December 2027 ends with a check against a coverage target of 85 out of 100 for the target role, and anything still open is re-sequenced into spring.
A step drops off only when the prompt behind it is answered. Finishing a course or a case study is not the same as recording it, so the evidence page asks for the details (the audience size at the showcase, for example) before coverage moves.
What does the roadmap not do?
It does not do the work for you or promise an offer. It tells you which gap to close next and why, and the evidence still comes from you: the course you take, the case study you write, the talk you give.
It is also not a template to copy. Every name and number in this sample is invented, down to the coverage target of 85; your own roadmap is built from the gaps on your own evidence page.
How is this different from a generic twelve-month plan?
A generic plan lists good ideas in no particular order. This one is ordered by what depends on what: the course comes before the case study that uses it, the case study comes before the January applications that link to it, and the internship comes before the senior-year case studies built from it.
It also follows the calendar. Applications sit in the month their deadline sets, winter break carries a single task, and summer is left to the internship. A plan that ignored the terms would put applications over a break and new coursework in the middle of recruiting season.
How does the roadmap connect to the rest of the dashboard?
It reads from the other pages and sends work back to them. The two September messages come from the network page, and the referrals from Priya and Marcus a year later are the same warm paths paying off. The case studies land in the portfolio, and the December interview uses an answer the interview page drafted from your evidence.
The two buttons above show the loop. Update the evidence and a step can drop off; change the target career on the careers page and the whole order rebuilds around the new role. Clean-up sits apart: it lets you flag things privately, and it never touches the AI-exposure number.
Which plan includes the roadmap?
The full term-by-term roadmap is part of Premium, together with a resume and LinkedIn that update as you add experience and a plan for who to talk to next. The current prices are in the pricing section of the home page.
The Free tier gives you the AI-exposure score, the map of durable and exposed tasks, and median salary context, which is enough to choose a target role before you plan toward it. Beta users get the first year free after launch; request beta access to be told when your account is ready.
Method and sources
The published AI-exposure number is built only from O*NET task data, Eloundou et al., and METR*. Pay, growth, networking, portfolio, and clean-up are context and never move it. A 0–100 exposure scale.
Written by Brooke Houck, PhD.
This dashboard is a sample. Maya Rivera, Northlight University, Northbeam, Lumen, and Verdi Labs are invented, and every figure on this page is sample data, not a real result.
* This projection uses METR's established long-run time-horizon trend, not the faster 2026 estimates. Those newer figures come from a task suite near saturation, carry very wide confidence intervals, and measure software tasks specifically, so extrapolating them across all occupations is not yet reliable. See MIT Technology Review, "This is the most misunderstood graph in AI," February 5, 2026, https://www.technologyreview.com/2026/02/05/1132254/this-is-the-most-misunderstood-graph-in-ai/