Tier · Pulse₹2L / month · Monthly, 3-month minimum
Pulse
Production AI observability.
Pulse is always-on observability for AI systems in production. Cost, latency, drift, hallucination rate, user-reported failures — all tracked, with a weekly digest and an escalation path when something breaks at 2am.
Who it's for
Teams with at least one AI feature in production who can't afford for it to silently degrade, but can't justify a full-time ML reliability hire.
What you leave with
- Evals + monitoring pipeline installed and owned by us
- Weekly digest email to engineering leadership
- Pager rotation for P1 drift / outage events
- Monthly written trend memo + optimization recommendations
How it runs
- Month 0Setup: telemetry wiring, eval set creation, SLO definition.
- OngoingWeekly digest. Monthly memo. P1 pager response.
Explicitly out of scope
- Root-cause fixes beyond configuration (those convert to Forge tickets)
- User-facing support (your team owns user comms; we own engineering escalation)
Next tier
Spark — One problem. Two weeks. Prototype + memo.