Uma Mendez

Senior Engineer · ML Platform · Data & ML

ActiveActive 1d ago

Craft posture

-5 vs last month

Tokens used / mo

74M tokens

74,279,863 tokens · spend, not output

Optimizable / mo (est.)

15M tokens

20% of tokens recoverable

Acceptance rate

76%in band

19 sessions / wk · band ~55–80%

Craft posture vs team

62craft posture

-5vs last month

Where they're strong and where to coach — PROXY scores from telemetry; weights in Settings → Scoring

Prompt clarity49

Planning / plan mode68

Test-driven flow49

Context management71

Tool selection47

Diff review discipline74

Security hygiene75

Prioritized by value at stake

medium

Right tool for the job — Tool selection

Scores 47/100 on Tool selection, the lowest in their profile.

Action: Review when to use search vs. edit vs. subagentsExpected: Faster task completion

Your 1:1 history

No notes yet. Add one after your next 1:1.