Starting an independent practice
Make the first-client mistakes here, with nothing at stake. Interview eight people, dig through realistic data, and ship a fix before a paying client is relying on you.
Start free →A forward-deployed engineer simulator
Free for aspiring FDEs. No credit card. Whether you are starting an independent practice or looking to land an FDE role, run a realistic engagement for a fictional company: interview the staff, find the real problem, build the fix inside their systems, and measure the result.
No sign-up to start. This is a full engagement, not a quiz. The simulated clock runs a week in about ten minutes, but the real work (interviews, data, a process map, a build) takes much longer, and your progress is saved so you can stop and come back. It works on a phone, but a desktop is much more comfortable.
Make the first-client mistakes here, with nothing at stake. Interview eight people, dig through realistic data, and ship a fix before a paying client is relying on you.
Start free →Your scores, interview transcripts, and the feedback from the people you interviewed become evidence that employers can review, so your work speaks before your résumé does.
Start free →Agent evals
We ran the same engagement humans take through AI agents, scored by the same rubric. Best so far: 92.4 out of 100. Nobody got it all.
A local plumbing and HVAC company is straining under its own demand. Customers sit on hold, Monday mornings are chaos, and the COO is ready to hire more people or hand the phones to AI.
“Maybe we hire two more, maybe AI answers the phone. You tell me.” Dana Okafor, COO
Your job is to find out what is really going wrong before anyone spends a dollar on the wrong fix.
The same loop a real engagement follows: understand the work, measure it, change it, and measure again.
Eight characters share some facts freely. The rest only come out when you ask a specific, well-aimed question. Talk about replacing staff and they stop talking.
A simulated company generates calls, jobs and messages across four weeks. Browse it in the app, filter it, trace one job end to end, and export it as CSV.
One simulated week runs in about ten real minutes. Week one is always the baseline, and a deploy takes effect the following week.
A rubric scores every phase, names its penalties, and includes how the people you interviewed felt about working with you.
Measure and readout
After you deploy, the simulation runs forward and shows what actually changed. Your baseline is the yardstick, so a weak baseline shows up here. So does a fix that costs more than it saves.
| KPI | Your baseline | Week 2 | Week 3 | Week 4 |
|---|---|---|---|---|
| Median turnaround | 6.2 days | 2.8 | 2.1 | 1.9 |
| Cost per case | $14.10 | $9.80 | $7.40 | $6.90 |
| Cases handled first time | 58% | 66% | 74% | 79% |
| Customer complaints | 41 | 33 | 24 | 19 |
| Run cost of the new system | $0 | $420 | $610 | $640 |
Example data, not from a real scenario. Better is green, worse is red.
Level 1 · Scenario 1
Tidewell Home Services books plumbing and HVAC work for 52 technicians. The phones are busy, customers wait, and the sponsor has two fixes in mind. Neither may address the real problem.
You will not be told the answer. You will be told who to talk to.
Start Level 1Guides
Short guides to the job, the interviews, and the methods the simulator scores.
All guides →