metrics / launch
Growth and measurement systems
Translating ambiguous growth goals into metrics, experiments, launch criteria, and operating feedback.
Projects
Project threads across measurement, recommendation, ads, causal inference, and applied AI.
Technical work organized by problem class, system boundary, tradeoffs, and reusable artifacts.
Project directions
These directions connect writing, system diagrams, and future public technical artifacts.
ranking / ads
Recommendation and ads infrastructure
Ranking, retrieval, allocation, and measurement layers that balance user value, marketplace constraints, and business goals.
holdout / lift
Causal measurement and experimentation
Separating lift from correlation with holdouts, attribution windows, and counterfactual reasoning.
agent / review
Enterprise AI and agent workflows
Integrating agents into real workflows through evaluation, escalation, human review, and measurable operating contracts.
Project note structure
Each project note keeps the problem, boundary, tradeoffs, artifact, and transferable lesson visible.
01
Problem shape
Describe the class of decision problem, not the private organization.
02
System boundary
Explain inputs, outputs, constraints, and feedback signals.
03
Tradeoffs
Speed vs validity, automation vs review, growth vs guardrails.
04
Reusable artifact
Link a diagram, checklist, note, or repo when it is public.
05
Transferable lesson
End with a transferable lesson, not an inflated performance claim.
boundary
Employer or client identifiers
Replace with domain-level descriptions: enterprise AI, recommendation systems, ads platforms, experiment infrastructure.
system map
Private metrics and internal architecture
Keep qualitative system goals, public diagrams, and general constraints. Leave out non-public details.
claim check
Inflated outcome claims
Do not invent numbers. If a number is not public and attributable, leave it out.