Exam integrity over everything.
A live session cannot be interrupted by an error, a loading state, or a connectivity warning. Every decision about state, errors, and async work is subordinate to this rule.
Who's behind it
Built by someone who's graded the run.
Paret is built by Asher Rommel — a paramedic and first aid instructor from Squamish, BC, who got tired of the paperwork getting in the way of running a good exam.
Origin
The form is the bottleneck.
The problem showed up the same way every time. Running scenarios for paramedic students in BC, I was constantly bouncing between three things at once: the scenario sheet to give the student their cues, the notepad to record what they did, and the student themselves — the only thing I was actually supposed to be watching. On practice runs it was manageable. On actual exam days — needing to observe, inform, and document everything accurately, all at once — it wasn't.
I started building Paret the day I realized this wasn't a personal failing. It was a design failure. Every examiner was solving the same problem with a clipboard and a good memory. The watching was the job. The writing was getting in the way of doing it well.
The fix I wanted was simple, even if the engineering wasn't: the examiner watches; the app writes; the candidate gets the same quality of assessment whether the exam runs at 9am or 5pm. And there was no good reason to limit it to EMS — the same problem exists anywhere a candidate is graded live against a checklist. So I didn't.
How we build01 02 03 04 05 06
Six rules we hold to, even when they're inconvenient.
Precision earns trust.
The interface doesn’t charm — it’s exactly right. Tap targets sized for gloved hands. Instant response. Labels that say precisely what the action does. No "are you sure?" loops for examiners who already know.
Roles determine the experience.
The examiner app and the student app are different products that happen to share a codebase. No screen tries to serve both. Every layout reads the role and assumes one audience.
Show, don't navigate.
The interface surfaces what the examiner needs next based on what the candidate is doing — scenario context, pending notes, the relevant section of the active sheet — without making them tap their way to it. Eyes stay on the candidate.
Institutional weight, not enterprise wallpaper.
Paret looks like it belongs in a training program's official toolkit, alongside paper forms and assessment rubrics. Not polished-startup. Not government-portal. The seriousness of the work shows in the design.
The AI doesn't have a personality.
Notes are written the way an experienced examiner would write them: past tense, third person, no judgment. No "I noticed…" No emoji. No celebratory micro-copy. The examiner decides what counts as correct.
Where we are
A short, honest timeline.
Q1–Q2 2026 · Live
First prototype
Live transcription proven against real exam audio. Note drafts are concise, past tense, and examiner-voiced.
Q3–Q4 2026
Closed pilots
Two BC EMS programs running EMR / PCP practical exams against the early build.
H2 2026
Waitlist opens
Programs across EMS, fire, police, and trades onboarded in cohorts as forms are translated.
2027
General availability
Examiner + Student apps in production. Institution dashboards for program directors.
The founder
Someone who's stood at the foot of the manikin.
Paret is a solo project for now. Examiner advisors across EMS, fire, and trades are part of the build process — they weigh in on every decision that touches how exams are run.
Photo of Asher
Get in early
If you grade live, we want to hear from you.
Tell us what you assess and what forms you use. We're building the launch cohort around programs that reach out early.