Assessments — Index & Grading Guide
The graded learning layer for uEngage Prism. Five difficulty levels, each testing a different skill. Use alongside the New-Joiner Learning Path — each week's checkpoint points here.
The five difficulty levels
| Level |
Name |
Tests |
Format |
Where |
| L1 |
MCQs |
Recall of facts (what/where/which) |
Multiple choice, one correct + explanation |
Getting-Started, Business |
| L2 |
Scenarios |
Applying facts to a described situation |
Short-answer scenario |
Getting-Started, Business |
| L3 |
Code-reading |
Reading real files/functions and predicting behavior |
"What does this do / where would you look" |
Interview bank §6–7 |
| L4 |
Architecture |
System-level reasoning & trade-offs |
Design/why questions with model answers |
Level-4-Architecture |
| L5 |
Production-debugging |
Incident triage: symptom → investigation → root cause → fix |
Structured incident exercises |
Level-5-Production-Debugging |
Plus hands-on: Assignments — practical tasks (trace an order, add an endpoint, fix E1.9, add an index, feature flag, cron).
flowchart LR
L1[L1 MCQ<br/>recall] --> L2[L2 Scenario<br/>apply]
L2 --> L3[L3 Code-read<br/>predict]
L3 --> L4[L4 Architecture<br/>reason]
L4 --> L5[L5 Debugging<br/>diagnose]
L5 --> A[Assignments<br/>build]
Assessment files
| File |
Level(s) |
Topic |
| 00-Getting-Started-Assessment.md |
L1 + L2 |
Setup, system overview, deployables, stores, auth |
| 01-Business-Assessment.md |
L1 + L2 |
Product/business model, RFM, journeys, loyalty, membership, revenue |
| Level-4-Architecture.md |
L4 |
Dual repos, crm_queue seam, Kafka, dual stores, dual dashboards / 3 pipelines, non-atomic points, in-process crons |
| Level-5-Production-Debugging.md |
L5 |
6 realistic incidents (order missing, E1.10 points, campaign not delivered, E1.9 AOV=0, cron stuck, E1.3 WhatsApp volume) |
| Assignments.md |
Hands-on |
Trace order, new endpoint, DB field, fix bug, index, feature flag, cron |
Scoring & self-grading guide
Per-level scoring
- L1 MCQ: 1 point per correct answer. No partial credit. Pass = 80%.
- L2 scenario: score each 0 / 1 / 2 — 0 wrong, 1 right conclusion but weak reasoning, 2 right conclusion with the correct mechanism/collection/cron named. Pass = average ≥ 1.5.
- L3 code-reading: 0/1/2 — 2 requires naming the exact file/function and the behavior. Pass = ≥ 1.5.
- L4 architecture: compare against the model answer; award 2 for capturing the key trade-off/risk, 1 for a partial answer. Pass = ≥ 75% of max.
- L5 debugging: each incident scored on 4 axes (symptom understood / correct investigation path / correct root cause / viable fix), 1 point each = 4 max per incident. Pass = ≥ 3/4 average across attempted incidents.
Overall competency mapping
| Overall score |
Level |
Meaning |
| < 50% |
Onboarding |
Keep working through the Learning Path |
| 50–79% |
Contributing |
Can take supervised tickets |
| 80–90% |
Independent |
Can own module-scoped work + on-call shadow |
| > 90% |
Fluent |
Can lead debugging / review others |
Self-grading tips
- Don't peek. Write your answer first, then reveal.
- For L2–L5, a right answer that can't name the collection, cron, or file is only half-right — Prism rewards specificity.
- Re-take any level you fail after re-reading the linked KB page. Track weak areas in the Learning Path tracker.