Department

AI Engineering Lab

30 engineers · 4 teams

Operating lift

1.73/ 2

+0.10 QoQ

AI adoption

99%

of engineers

$ / seat · mo

$296

Trained

83%

25/30 engineers

Top outcome

32% fewer bugs reaching prod since the trained cohort ramped.

Teams in this departmentClick a team

Agent RuntimeLead · Priya Raman

1.8199%

Model EvaluationLead · Wei-Lin Ong

1.6499%

Eval InfrastructureLead · Dev Kapoor

1.5899%

Applied ResearchLead · Sofia Marchetti

1.4296%

Recommended actions

AI Engineering LabExpand rollout

Agentic adoption is compounding — eval hooks and reasoning checkpoints are default, holding the department at 1.73.

Applied ResearchCoach cohort

Applied Research trails the lab — its agent-mode use is shallow and pulls the department average down.

All recommendations

How AI Engineering Lab uses AI

Full tasks

AI completes multi-step work end-to-end

48%

Code suggestions

inline help while engineers type

32%

Q&A

asking the AI questions

13%

Review help

AI checks code before it ships

Training liftJan → Jun 2026

PRs using AI agents95% · was 51%

Reviews with AI notes91% · was 59%

Runs with context summary84% · was 55%

AI PRs with eval hooks77% · was 39%

SpendSpend page

$ / issue

$104

10%

$ / seat · mo

$296

$ / 1k lines

$19

Hours / issue

5.4

10%

Budget used · month55% of $16.0k