Review Board — DevFlow Conf 2027

Events›DevFlow Conf 2027 — Moscone West, San Francisco, CA · May 12–14, 2027›Review Board

Public sites: cfp ↗program ↗signed in as swyx

Curtain Call does not use weighted scorecards or workflow-driven review rounds. Reviewers give signed Stars and comments; the committee converges through conversation and progressively narrowing the pool. Every submission stays on the shared table—no pagination, no page reloads, and no hiding what your fellow reviewers are doing.

Multi-round review: Statuses are the rounds: round one is the Pending queue, advancing a talk IS moving it to Accept Queue or Waitlist, and round two is the same table re-sorted over the narrowed pool. Scope each member to tracks for pools; the coverage sort (fewest ratings first) distributes work by pull, and the 2-review bar shows when a pass is saturated. → the board's two sorts · per-member track scope
Weighted multi-criterion scorecards: One star, many reasons — and the data is yours: you can even turn on additional rating fields here (feature-flagged, off by default) and export them through the API — then weight them in your own spreadsheet. We refuse the reporting layer, not your analysis. → Exports & API
Blind review as a global mode: Blind is one switch with a live reveal: flip it at the phase boundary for a blind first pass and an open second — the outcome of the per-round matrix without the matrix. → the visibility switch

The full list of what we refuse, and why: the Manifesto.

Status:Accepted 4 Pending 3

COVERAGE

Coverage worklist

# ratings ascending

2-review rule: rate the least-covered submissions first.

0% reviewed

0/7 have ≥2 reviews

DECISION

Decision queue

Mean stars descending

Compare mean stars directly; Reviews shows confidence and warns below the 2-review target. Submissions nobody has rated yet have no mean and sit at the end.

2 of 7 have a score

DECISION

Decision queue

Mean stars descending

Compare mean stars directly; Reviews shows confidence and warns below the 2-review target. Submissions nobody has rated yet have no mean and sit at the end.

2 of 7 have a score

7 submissions·0 fully reviewed (0%)·no close date set

(2 reviews · 3 comments)

Aug 19Oct 2

7 talks need coverage before deciding · target 2 reviews each

Track:AI Engineering 2 Platform & Infra 1 Developer Experience 1 no track 3

Coverage worklist · fewest ratings first · the 5 nobody has rated lead.
Speaker Fname · Lname · Org	Format	Stars	Reviews	You
Ravi Teja Prabhala Venkata revealed to you	Talk (30 min)	—	0 ⚠	–	Read & rate → PendingQuick rate ▾
Senior Engineering Manager · Capital One
The Agent Harness Nobody Writes About: Permissions, Second-Model Review, Evals at Volume AI Engineering
Marcus Okafor revealed to you	Lightning Talk (10 min)	—	0 ⚠	–	Read → Accepted
Staff Developer Advocate · Cloudreach Labs
Lightning: Agents in Production Q&A AI Engineering
Priya Raman revealed to you	Talk (30 min)	—	0 ⚠	–	Read & rate → PendingQuick rate ▾
Principal Engineer · Latticework Systems
QA-PROBE-SPK-S2: account linking test (please ignore) Platform & Infra
Priya Raman revealed to you	Lightning Talk (10 min)	—	0 ⚠	–	Read → Accepted
Principal Engineer · Latticework Systems
Docs That Answer Back: Retrieval-Grounded Documentation Sites Developer Experience
Priya Raman revealed to you	Talk	—	0 ⚠	–	Read & rate → PendingQuick rate ▾
Principal Engineer · Latticework Systems
Docs That Answer Back: Retrieval-Grounded Documentation Sites
Priya Raman revealed to you	Talk	5.0	1 ⚠	–	Read → Accepted
Principal Engineer · Latticework Systems
Your AI Pair Programmer Is Lying to You: Verification Patterns That Scale
12345 ★5Sam Whitfield — Originality 5/5, Relevance 5/5. Excellent fit for the AI Engineering track.
Priya Raman revealed to you	Talk	4.0	1 ⚠	–	Read → Accepted
Principal Engineer · Latticework Systems
UPDATED: Taming 40-Minute CI: Incremental Builds at Monorepo Scale Speakers: Priya Raman (Primary speaker) · Marcus Okafor (Co-speaker)
12345 ★4Sam Whitfield — Scoring note (multi-criterion breakdown, since this board uses a single signed star rating): Originality 4/5, Relevance 2/5 — unweighted average 3.0, weighted average (Originality x2) ~3.33; rated 3.5 stars overall. Recommendation: Accept. Strong practical content and a clear narrative arc; abstract could name the specific tooling used. Recommend accept for the Platform track. Sam Whitfield — Strong practical content and a clear narrative arc; abstract could name the specific tooling used. Recommend accept for the Platform track.

2 of 7 rated · 5 unrated · 7 of 7 shown · Ratings are 1–5 with halves; SPLIT marks a spread of 2.0 stars or more — the rows worth arguing about on the call.

Build dd0cf13 · deployed Sep 28, 2026 at 21:12:04 UTC (3d ago)