Monitoring
monitoring
AI Engineering
Separate retrieval, packing, and generation failures
Use a request trace to tell retrieval, context-packing, and generation failures apart, and report answer quality with a named denominator next to operational signals.
Design a release gate with slices and uncertainty
Decide whether a candidate with better aggregate success but a regressed protected slice can ship, and design a release gate with calibrated graders, traces, and stated uncertainty.
Design time-aware evaluation and drift response
Design rolling-origin backtests, capacity-bound metrics, label-delay handling, and drift monitoring for a synthetic card-fraud model whose chargebacks arrive weeks late, then decide whether to promote, shadow, or hold a retrained candidate.
Community
Metrics, Logs, and Traces: The Three Pillars Without the Marketing
What each pillar actually does, when reaching for it pays off, and the budget I follow so I am not paying observability vendors more than I am paying for compute.
Datadog Onsite: Five Hours of System Design
A Datadog senior backend onsite where four of the five rounds were system design, anchored on real telemetry-shaped problems.
DevOps / SRE Interview: The Production Postmortem Round
An SRE loop at a Series D infra company, anchored on the round where they handed me a real-feeling postmortem and asked me to find what was missing.
SRE Incident Drill Questions From Our On-Call
Four incident-shaped questions our on-call rotation uses to interview SREs. Each one starts with a symptom, asks you to write the smallest diagnostic snippet, and the discussion is more important than the code.
Core Web Vitals Without the Buzzwords
LCP, CLS, INP measure three real user-felt experiences: load, stability, responsiveness. Treat them as field truth, not Lighthouse scores to game.
