ADVISORY
Charles Trinh — Architecture Review Board Design | Las Vegas, NV · Las Vegas, Nevada · Henderson · Clark County, Nevada · North Las Vegas, Nevada

An architecture review board only earns its meeting slot if it changes decisions rather than records them. I design the board so it makes a small number of binding calls a year, each written down with the reasoning that produced it.
A board that reviews everything decides nothing
Most review boards I am asked to rescue meet monthly, receive a deck, and produce polite comments. Nobody is bound by the outcome, so the same argument returns two months later. The board's membership is often a roll-up of every senior engineer, which means no one has the authority to veto. My first job is to cut the remit down to three or four classes of decision that genuinely need cross-team agreement: shared data contracts, choice of vendor for anything the whole organisation depends on, changes to authentication or network boundaries, and anything that spends more than a stated threshold.
Everything below that line is delegated in writing to a team lead, who keeps their own record. The cut is the whole design. If the board sits above every pull request it becomes a bottleneck and the teams route around it; if it sits above nothing, it is theatre. The artefact that settles this is a two-page charter listing the decision classes, who owns each, and how a team appeals. I draft it with the CTO and the head of security in the room, because those two are the only people who can grant the authority it needs.
Reviewers are appointed for twelve months by name, not by job title. A title-led board silently changes shape when someone leaves and a successor is never nominated. Named reviewers are easier to call to account, and easier to remove when they start blocking work for reasons that have nothing to do with the stated classes. The cost of a badly appointed board is measurable: teams stop bringing early designs, rework lands late, and the first review a new system gets is the one after the incident.
Every decision leaves a record someone can reverse
The durable output of a review board is not its minutes. It is the architecture decision record, a short file held in the same repository as the system it governs. Each record names the decision, the date, the reviewers who approved it, the alternatives they considered, and the conditions under which the decision should be reopened. That last field is the one teams skip, and it is the one that stops a choice from hardening into folklore. When someone later proposes a different database model, the record tells them what the original reviewers knew.
I write the first twenty records myself so the format is set by example rather than by instruction. After that the board writes them, with a template and a naming convention that makes them searchable. A record that takes more than a page is usually two decisions tangled together, so the board splits it. I keep a monthly index of open decisions and their reversal conditions, and I bring it to the first review of each new quarter. It is surprising how many of them should already have been reversed by then.
The record also carries the weight of the meeting. If a decision cannot be expressed as a record with named alternatives, it was not a decision, it was a preference. Sending such items back to the proposing team costs one awkward conversation and saves the board a month of re-litigation. Over time the board stops receiving preferences and starts receiving well-formed choices, which is the point where reviewers begin turning up prepared. I watch for that shift because it is the most reliable sign the board is working.
Quorum, minutes and the meeting that should not exist
A board without quorum rules gets hijacked by whoever turns up. I set a written minimum: the chair, the security lead and at least one engineering reviewer outside the proposing team. If quorum fails, the meeting is cancelled, not held with substitutes. The chair holds a casting veto but has to write a record explaining it within two working days, which keeps that power visible. Minutes are one line per item and a link to the record; they are not a transcript and nobody should need to write them afterwards.
Before each meeting I review the incoming list with the chair and ask one question of every item: does this still need the board, or can a single owner be named and the decision delegated with a diary date to report back? Roughly a third of items fail that test. The board does not accept them. This triage is the single highest-leverage thing the chair does, and it is almost never written into anyone's job description, so I put it in the charter explicitly with a time budget of forty minutes a week.
For an organisation setting up its first board, a short weekly office hour with the fractional architect is usually enough for the first quarter, provided the charter and the record template are fixed up front. For a larger estate, the record format should also line up with the platform team's runbooks so there is one place to look for why a system is shaped the way it is. The charter, the template and the first twenty records are the deliverable.
Membership is a rotating seat, not a career
A fixed board composed of the same people for years develops loyalties to its own past decisions. I rotate one engineering seat every six months and keep the chair, the security reviewer and one platform reviewer for continuity. New members are briefed on the open records and the reversal conditions before their first meeting, and they are explicitly expected to challenge old choices where the conditions have been met. That brief takes an hour and stops the drift where new members spend three meetings learning the room before they say anything.
The chair is accountable for throughput, not for consensus. If the board cannot decide in the meeting, the chair records why, names the missing input and sets a date to return. An item that returns without the input is closed and the team is redirected to the authority who can supply it. This sounds harsh and it is the only mechanism I have found that stops the undecidable items accumulating into a permanent backlog that everyone learns to ignore. I track the median time from submission to decision, and the target is five working days.
The cost of getting this wrong is not just slow decisions. When a board is vague, engineers make architectural choices in code review at midnight and the organisation discovers them during an outage. A tight charter, a named reviewer list and a visible record keep those choices where they can be tested. My own delivery record inside MyRxWallet includes a 337-table clinical store governed under the same decision-record discipline, which is where some of these patterns were pressure-tested rather than merely described.
Connecting the board to the rest of the delivery machinery
The board cannot sit alone. Its decisions have to reach into the release process and the change record. I work with the platform team to place a record check at the point where a service is promoted between environments, and I pair that with the release gate design so a decision that has been superseded cannot quietly ship. This is a small script and a naming convention, not a new system. The aim is that the review trail exists where the release actually happens.
Security decisions get a dedicated reviewer who can attend a subset of meetings rather than every one. Where the organisation is moving toward HIPAA attestation or a cloud architecture review, the technical decisions and the compliance evidence draw on the same records, and I keep that link visible in the board's index. The HIPAA security programme and the board should look at the same evidence with different eyes, not maintain two parallel paper trails that disagree.
When the board is running, I stay for two quarters and then step back to an external review of the records twice a year. That is enough to catch drift without becoming a standing member nobody can remove. The handover pack is the charter, the reviewer roster, the index and the last quarter's decisions with their owners and dates. If the board cannot run without me in the room, the design failed, and I say so at the handover review.
Proof
- ONC (g)(10) self-test PASS, 317 of 317 subtests, April 2026 — public suite URL
- 67 public hostnames, 51 serving, one AWS account
- 104 clearinghouse enrolments across 53 jurisdictions
- USCDI v7 proposal — Patient Digital Identity Credential — listed at isp.healthit.gov
- SAM.gov UEI MVKQW9DKKGB8 · CAGE 20KM5 · NPI 1194645564
- AWS Business Associate Addendum active since 18 August 2026
Where this work is delivered
Engagements run from Las Vegas and reach across the valley: clients in Las Vegas, NV; Las Vegas, Nevada; Henderson; Clark County, Nevada; and North Las Vegas, Nevada — plus remote delivery nationwide.
How an engagement runs
1. Trace one real case
Follow a single patient or claim end to end with a timestamp at every hop, before any document is written.
2. Written findings
A findings list naming the defect, the system, the owner and the date.
3. Monthly review
Review the list monthly and check whether the dates held.
Frequently asked
What is this service?
This page describes charles trinh — architecture review board design — the scope, the failure modes it prevents and how the work is measured. It sits in the Security and Compliance category at trinh.io.
Where is the work delivered?
Engagements run from Las Vegas, Nevada across the valley — Las Vegas, Henderson, North Las Vegas, Spring Valley, Paradise, Winchester, Sunrise Manor, Enterprise, Summerlin, Clark County — and remotely anywhere in the United States. On-site work in the Las Vegas valley, remote delivery elsewhere.
What does it cost?
Standard rate $195 per hour, committed $165 per hour at twenty hours or more per week for three months or more; $1,600 standard day, $1,350 committed day. Scoped in writing before work starts, fixed fee where the deliverable is defined.
Who is Charles Trinh?
A technology architect in Las Vegas, Nevada, working in healthcare identity, consent and interoperability. Not a physician, not a medical practice, and not affiliated with any other Charles Trinh.
How does an engagement start?
With one trace of a real case and a written findings list naming the defect, the system, the owner and the date. The first call is twenty minutes and free: 702-559-3489.
Engagement
Scoped in writing before work starts; fixed fee where the deliverable is defined. First call is 20 minutes and free. 702-559-3489 · charles@trinh.io
Talk to Charles
- Phone: 702-559-3489
- Email: charles@trinh.io
- Based in Las Vegas, Nevada · working remotely across the United States
Record
Profile
Healthcare technology architect working where federal regulation becomes production infrastructure. Built and operates a FHIR R4 clinical platform — SMART App Launch authorization server, patient and provider portals, SQLite — on a hardened single-host AWS estate of 67 public hostnames. The platform returned a self-test PASS on the ONC (g)(10) Standardized API test suite in April 2026, 317 of 317 subtests, publicly verifiable. Designed a consent architecture in which a permissioned ledger records patient grant, refusal and revocation as immutable events while all clinical data stays off-chain and encrypted per patient — no PHI on the ledger by design. Filed a proposed USCDI v7 data element with ONC on verifiable patient identity. Translates the Federal Register, HHS/CMS guidance and HITRUST requirements into architecture, controls and audit evidence.
Selected engineering scale
- Self-managed estate of 67 public hostnames, 51 serving, on one AWS account.
- FHIR R4 API with a SMART App Launch authorization server; patient and provider portals; an AI service layer; Hyperledger Fabric peer, orderer and CA; IPFS; a 337-table SQLite clinical store.
- Two permissioned distributed ledgers in production, one operating a consent registry that records grant, refusal and revocation with no PHI on chain.
- X12 005010 pipelines across eight transaction sets.
- Designed, built, deployed and operated end to end by one person.
Credentials and standards work
- Inferno (g)(10) self-test PASS, 317 of 317 subtests, April 2026 — the ONC Standardized API test suite. Standards validated: US Core 6.1.0, USCDI v3, SMART App Launch 2.0.0, Bulk Data 2.0.0. Public suite URL.
- USCDI v7 public comment submission, April 2026 — proposed Patient Digital Identity Credential under the Provenance data class, listed publicly on the ONC Interoperability Standards Platform.
- SAM.gov UEI MVKQW9DKKGB8 · CAGE 20KM5 · NPI 1194645564. AWS Business Associate Addendum active 18 August 2026.
Technical environment
AWS us-east-1 (compute, storage, messaging, SMS, email, voice) · Linux and systemd hardening · nginx · Fabric and EVM ledgers · FHIR R4 · X12 005010 · HL7 · SQLite · self-hosted mail, drive, search and monitoring · Git-based change control and CI release gates.