Skip to content
Cinnabar / Systems

Enterprise platforms. Heavy bench.

Systems is the practice for engagements where the technical complexity is non-negotiable and the stakes are visible from the C-suite. ERP-class platforms, vertical industry systems, crypto and financial infrastructure — the work that has to be right before it ships.

What Systems is

The work that has to be right before it ships.

Systems is the practice we run for partners whose technical complexity isn't something to engineer around — it's the substance of the business. Vertical operators competing against incumbent platforms, financial-infrastructure teams, crypto custody, ERP-class operations.

These are engagements where a single bug is material; where the audit posture is regulated; where downtime is measured in dollars per minute and visible from the C-suite. The right answer is rarely the most novel one — it's the most boring choice that satisfies the constraints, executed with care.

Systems engagements run twelve to twenty-four months with a senior bench. Every contributor has shipped a real regulated, high-stakes, or high-volume system before. We don't do staff augmentation, we don't outsource execution, and we don't put juniors on partner code.

Services

What a Systems engagement covers.

  • ERP & operations platforms

    ERP-class systems for operators that have outgrown off-the-shelf — finance, supply chain, scheduling, settlement. Built around the workflows your team already trusts, not a vendor's demo flow.

  • Vertical industry systems

    Vertical platforms for operators competing against horizontal incumbents — real estate, healthcare, logistics, insurance. Built so the company's defensible asset (its listings, providers, drivers, policies) lives in a database your team controls.

  • Crypto & settlement infrastructure

    End-to-end custody, settlement, and reconciliation for stable-coin issuers, exchanges, and on-chain treasuries. HSM-rooted signing, threshold approvals, real-time reconciliation against off-chain ledgers.

  • Financial platforms

    Ledger systems, payments rails, treasury operations, and the audit trail behind them. Idempotent at every layer; double-entry where it matters; reconciliation as a first-class feature, not a quarterly script.

  • Legacy replacement

    Replatforming twenty-year-old systems without customer-visible downtime. Parallel architecture, shadow-mode validation, customer-by-customer cutover, and the operational scaffolding the legacy system never had.

  • Compliance-grade infrastructure

    SOC 2 Type II, HIPAA, PCI, and the audit trail each one demands. Encryption-at-rest with per-tenant keys, audit logging on every read of sensitive data, BAA-covered infrastructure, and a security posture that survives a real audit.

How we approach it

Five principles that shape every Systems engagement.

01

Constraints before architecture.

Systems engagements live or die on constraints — uptime, latency budget, regulatory regime, audit posture, the migration path from the existing system. We map every constraint before we sketch a single architecture, because the architecture is downstream of them.

02

Boring choices for load-bearing parts.

Postgres, well-understood queueing, well-understood compute. Novel choices reserved for the parts of the system where novelty earns its keep — the search index, the eval harness, the on-chain layer. Boring everywhere else.

03

Reconciliation as a first-class feature.

Every system that touches money or material reality needs continuous reconciliation against the world it represents. We build it as a first-class feature, not a quarterly batch job — discrepancies trigger an automatic freeze and a page, not a Monday-morning surprise.

04

Replatforms done quietly.

Parallel architecture, shadow-mode validation, customer-by-customer cutover. The first signal that a customer's system was replaced should be that it stopped breaking. There is no marketing event around the swap.

05

Audit trail as an operational tool.

The audit log isn't a compliance artifact — it's the single most useful operational tool the system has. Replay, blame, trend analysis, policy review. We build it to be queried by your team daily, not by your auditor annually.

Tools we reach for
  • TypeScript
  • Rust
  • Python
  • PostgreSQL
  • PostGIS
  • Elasticsearch
  • Kafka
  • Temporal
  • AWS Nitro Enclaves
  • YubiHSM
  • Foundry
  • Terraform
  • OpenTelemetry
How it starts

A constraints-mapping discovery, then an honest fit conversation.

Systems engagements don't start with a brief — they start with a constraint map. Uptime targets, latency budgets, regulatory regime, audit posture, the migration path from whatever exists today. We do this in a paid two- to four-week discovery, with two senior engineers and direct access to your operations and compliance leads.

The output is a real architecture proposal, an estimate we'll stand behind, and a clear answer on whether we're the right team to run the engagement.

Ready when you are.

A 30-minute conversation. We'll listen. If we're a fit, we'll say so. If not, we'll point you to someone who is.

No discovery decks · No sales calls · One conversation