← romesh.tech

ROMESH.TECH · ARCHITECTURE REVIEW

Before you commit to an expensive system decision, get an independent technical read.

Migrations, rewrites, scaling changes and architecture bets often look reasonable until hidden assumptions meet production.

I pressure-test the decision, expose the assumptions, and tell you what I would do and why.

Discuss your decision

YOU LEAVE WITH

  • What I think you should do
  • Why
  • What could fail
  • What I would test before committing
  • What can safely wait

USEFUL WHEN

You are about to migrate something important

Identity provider, payment system, database, infrastructure, vendor, regulated workflow.

The team disagrees on architecture

Rewrite vs incremental change. Build vs buy. Events vs polling. Scale now vs later.

The system works, but growth is making it dangerous

More customers, assets, integrations or operational load are increasing fragility.

The change is hard to reverse

You want someone outside the implementation loop to challenge the assumptions before commitment.

WHAT AN ARCHITECTURE REVIEW LOOKS LIKE

  1. You send the decision and constraints

    The problem, current constraints and the decision you are considering.

  2. I review the system and assumptions

    The architecture, assumptions and failure modes.

  3. You get a written decision memo

    • Recommended direction
    • Trade-offs
    • Failure modes
    • Assumptions that need testing
    • Reversible vs irreversible decisions
    • Immediate next steps

  4. We walk through the result

    If the decision needs deeper work, we decide that afterwards.

SELECTED WORK

Hours → ~200ms

High-volume settlement redesign

Reworked a high-volume settlement path and measured the result on the same clock.

Open the full chain ↗

~40,000 records

Controlled regulatory migration

Rehearsed and executed a regulator-driven migration with controlled reversal points and no onboarding downtime during cut-over.

Open the full chain ↗

30 → three-figure design target

Scaling without rewriting the money-moving path

Designed a staged scaling path without rewriting the money-moving path.

Open the full chain ↗
Inspect the evidence

Have a decision you are not comfortable making blind?

Tell me what you are changing.

Discuss it