Skip to main content

0001 — Second FE engineer for Bifrost (Syafrizal M. over-capacity)

Delivery / Program Management — decision record (ADR-style). Filed at the team level (bifrost/delivery/decisions/) rather than under a single initiative, since the underlying constraint — Syafrizal M. is Bifrost's only FE engineer — spans every initiative with FE scope, not one; capacity.md and roadmap.md are the canonical team-level artifacts this decision would update.

Context​

Syafrizal M. is the only FE engineer on Bifrost. As each 26Q3 initiative's task-breakdown was sized through 2026-07-02, every initiative with FE scope defaulted onto him with no alternative — by 2026-07-02 his known forward FE load peaked at 77 md (1.48× — a 25-day shortfall) before a 2026-07-02 centralized-sso-token effort re-baseline brought the canonical figure down to 63 md ÷ 52 effective = 1.21 (11 days over), per capacity.md and the roadmap's own closing reconciliation note (both are explicit that this — not the earlier 1.13 figure left as a dated historical entry — is the current number).

This is finding F-4 (off-plan-load) in the Sprint 1 reconciliation ledger (delivery/reconciliations/2026-07-06-sprint-01/), escalated across all 6 Sprint 1 reconciliation passes with no staffing action taken, and named again in Sprint 1's own retro action #4 ("Don't let this roll as a passive observation into Sprint 3"). Concretely, over-capacity blocks or degrades:

  • One CID Multiple WABA (P11) — FE 7.0 md has no queue slot at all; overflows to Q4 outright.
  • Centralized SSO Token (P12) — FE finish pushed to ~W10.6, past its own ~W10.3 Lane-B E2E gate need; already downprioritized P7→P12 partly for this reason.
  • Downgrade User (P10), Billing Expired Handling (P14), Postpaid Usage Scheduler (P16), Qontak Launchpad Audit Logs (P15) — each carries FE work "recorded on Syafrizal M." with no realistic schedule behind it.
  • 26Q3 Adhoc bucket (BIF-8695) — the staffed BE+FE stories route to ESCALATE rather than closing, purely because their FE half has no slot; as of 2026-07-19 that set is down to A3 (🔴 Expedite, customer-/revenue-facing) and A4 (🟡 Standard), meaning the bucket's own Pull rule is still flagging A3 as effectively FE-hire-gated, not schedule-gated. (A7 was resolved independently of this decision on 2026-07-19 — reclassed 🔴 Expedite and reassigned BE-only to Grehasta, since Modpanel needed no dedicated FE slot in the first place. A1 was resolved the same way the same day — its FE half is also the internal Modpanel surface, so it was reassigned BE-only to Satya. A5/A6 moved out to Self Top-up (P3) on 2026-07-16. None of the four contributes to this constraint anymore.)

The lever the docs consistently name is: add a second FE engineer, or move lower-priority FE scope off Syafrizal M. Neither has been decided. This ADR does not resolve which — it records that the decision is due and gives the EM/TPM a single place to land it.

Decision​

Pending — Bifrost EM/TPM to choose one (or a combination):

  1. Hire or reassign a second FE engineer onto Bifrost.
  2. De-scope or re-sequence lower-priority FE work off Syafrizal M. (which initiatives, and into which quarter) — an explicit roadmap ## Priority reorder or target_quarter slip would follow.
  3. Accept the overcommit as a recorded risk for a bounded period, with an explicit re-review date.

Record the choice here, update status: to approved, and propagate to capacity.md / roadmap.md and any affected initiative's README.md/timeline.md.

Consequences​

  • If a second FE engineer lands: re-run the sequencing across every FE-bearing initiative above (initiative-timeline-mapper territory) and re-pull the remaining ESCALATE-classed adhoc stories in Class order (A3 first per the bucket's Pull rule; A4 behind it).
  • If FE scope is de-scoped/re-sequenced instead: the affected initiatives' target_quarter and the roadmap ## Priority order need an explicit, dated edit — not a repeated sprint-doc deviation (Sprint 1 and Sprint 2 both ran Syafrizal M.'s queue in raw priority order as a "tactical, one-sprint-only" exception; per doctrine a deviation recurring a third time must be reconciled here or in the roadmap, not left as a sprint-doc footnote again).
  • Either way: this ADR is the artifact Sprint 1's retro action #4 asked for, and the record sprint-resync can point F-4 at until the EM/TPM decides.