↯Clarificeprototype
DayWeekQuarter
← Dashboard
RD

Roman Doyle

Engineer · Engineering · IC3 · reports to Bilal Toma

Useful score · Week ending Jul 7
60Solid
vs Engineering · IC3 (n=10)

Driving value

  • ▴ Tool leverage: top of cohort (85th pct)
  • ▴ Engagement (raw activity): top of cohort (95th pct)

Watch-outs

  • ▾ High activity but little delivered value (motion ≠ output)
Outcomes
0
docs, PRs, features, decisions
Value delivered
0.0
impact × credit
Days worked
1
no time off
Follow-through
100%
1/1 kept · 0 dropped

Activity flow

ClaudeSlackGitHub
2026-07-012026-07-01

Why this score

weighted sum of cohort percentiles
Value delivered0 pts/day · weight 45%50th pct → +22.5
Collaboration1 peers+threads · weight 20%60th pct → +12.0
Consistency0.5 0–1 · weight 15%50th pct → +7.5
Tool leverage3 tools · weight 10%85th pct → +8.5
Engagement (raw activity)4 signals/day · weight 10%95th pct → +9.5
Useful score60.0

Evidence vs level rubric

Senior IC · IC3
Evidence tracks at level

Evidence broadly supports the current-level expectations.

  • Impact & scope
    Owns features including their edge cases.
    ~Some evidence
    • · Value delivered in the 46th percentile of their track · 0 outcomes
    e.g. shipped the payments retry work end-to-end
  • Collaboration & influence
    Works well with their team and builds shared context.
    ~Some evidence
    • · Collaboration signal in the 45th percentile of track
  • Execution & reliability
    Reliably delivers what they commit to; unblocks others.
    ○Human-assessed
    • · Too few decided commitments this period to judge
  • Direction & judgment
    Makes sound technical/product decisions others build on.
    ✗No evidence
    • · No decisions or project leadership observed this period
  • Craft & quality
    Sustains high quality and a steady, dependable cadence.
    ~Some evidence
    • · Cadence in the 50th percentile of track

Evidence mapped to your imported rubric. A human makes the leveling call — this is the receipts, not the verdict. Evidence found for 4 of 5 expectations.

Follow-through

did they do what they committed to?
100%1/1 commitments kept
100% on time
  • ReviewReview PR: the unused importfrom Cora Whitfield · due 2026-07-01
    On time

AI leverage & augmentation

what agents can take on · what’s above an agent
0%of observed activity is routine enough to offload to an agent
Human-differentiated · 25%AI-assisted · 75%

Above an agent

  • ▴ Little human-differentiated output observed this period

Task-level augmentation analysis — where agents can help and where human work must be kept. Not a person-level replaceability score.

Value chains & cost

signal → meeting → doc → shipped · with a cost ledger

No reconstructed value chains in this period.

Outcomes this period

None recorded.