Founding Infrastructure Engineer

About Matterhaul

We're building the next-generation operating system for companies that build and distribute physical goods. The supply chain that moves industrial goods in this country still runs on 30-year-old software, manual re-entry, and email as the glue. Infor, Epicor, NetSuite, and SAP have had three decades to fix it and they haven't.

We're replacing all of that with AI. Matterhaul connects to the systems distributors already run, captures the operational context legacy software misses, and deploys agents that handle quoting, ordering, procurement, and dispatch — with humans in the loop on the things that matter. Customers are using it in production today, and we're in conversations with multi-billion-dollar distributors actively looking to move off their ERP.

Mission: remove friction from B2B supply for manufacturers and distributors, so every product flows reliably across the full order lifecycle.

Vision: a world where moving physical goods is as seamless as moving information.

Why now

Three things are true at once and rarely line up:

• AI has stopped being an experiment. A year ago this was a pilot conversation. Today, executives are being asked why they aren't deploying AI in operations.
• Legacy ERPs are no longer defensible. The customers know it. Their leadership teams are openly looking for a way out.
• The supply-chain trauma of the last five years is fresh. The companies that move physical goods lived through it in a way software companies didn't. They want visibility and automation, and they aren't going back.

The window is open. Windows like this don't stay open.

Why this role exists

Today we deploy to Railway: Docker-per-service, healthchecks, no Kubernetes, no IaC, shared Supabase Postgres, no data warehouse, no managed queue. It got us here. It will not get us into a distributor's critical workflow at scale, and it will not carry us through a Series A.

The next phase of the company is the platform — one cohesive system, not five tools that share a login. AI that shines. Workloads that customers can't operate without. That requires infrastructure built for it.

We're hiring a founding infrastructure engineer to rebuild our infra and core systems from zero: AWS + Kubernetes, clean service boundaries, Golang for core systems, and a deliberate set of choices around databases, data warehousing, and processing pipelines. You'll set the platform direction the rest of engineering builds on for years.

This is a high-leverage seat. You will not be re-skinning a Helm chart someone else wrote.

What you'll own

Year one, concretely:

• AWS landing zone — accounts, networking, IAM, KMS, secrets, CI/CD, cost controls. Codified in IaC (Terraform / Pulumi / CDK — your call, with rationale).
• Kubernetes platform — EKS or equivalent, with the supporting bits (ingress, service mesh if warranted, autoscaling, image pipeline, GitOps). Make boring choices that work.
• Service architecture — Carve crisp boundaries across what is today a mixed Golang + Node/Effect.ts service fleet (API gateway, GraphQL, workflow cluster, public catalog, durable streams, OpenFGA, telemetry, ElectricSQL proxy, healthcheck sidecar). Decide what stays, what merges, what gets rewritten. Push Golang as the language for core systems where it pays off — proxies, gateways, high-throughput pipelines, latency-sensitive paths — without forcing it where TypeScript/Effect is already a fit.
• Data layer strategy — We're on Supabase Postgres with Drizzle today, plus pgvector for embeddings, S3 for blobs, Redis for rate-limit state. You will make the call on:
  • Primary OLTP path forward (managed Postgres? Aurora? read-replica strategy? connection pooling? per-tenant isolation patterns?)
  • Data warehousing (Snowflake, BigQuery, Redshift, ClickHouse, Iceberg-on-S3, something else) — we have none today and need one.
  • Vector / search — keep pgvector, or move to a dedicated store?
  • Cache & ephemeral state — formal Redis/ElastiCache tier vs. in-process.
• Processing pipelines & async work — We currently run Effect.js workflow clusters plus a homegrown durable-streams service for everything async. It works but it's bespoke. You'll evaluate whether to harden it, augment it, or migrate selected workloads to off-the-shelf orchestration (Temporal, Airflow, Dagster, Flink, Kafka + Connect/Debezium for CDC, etc.) and build the ETL/ELT path into the warehouse from scratch.
• Observability — We have Sentry, OpenTelemetry, and Statsig wired. You'll extend with metrics/dashboards (Grafana / CloudWatch / Datadog — your call), SLOs, on-call rotation, and runbooks.
• Reliability & ops — Backups, DR, multi-AZ today and a credible multi-region story for tomorrow. Capacity planning. Cost ceilings. Security hardening (we already use OpenFGA for ReBAC and Clerk for authn — extend, don't unwind).
• Engineering leverage — Local dev that works in 10 minutes. CI that's fast. Deploy that's safe. Paved roads so product engineers don't have to think about infra.

You will write code. You will also make the architectural calls and document the "why" so the team can keep building after each decision.

What we're looking for

Must have:

• 7+ years building backend and infrastructure systems in production, including meaningful 0→1 platform work (not just operating someone else's platform).
• Production Golang — not just side-project Golang. You've shipped, debugged, and operated Golang services that real customers depend on.
• AWS at depth — VPC, IAM, RDS, S3, EKS/ECS, ALB/NLB, CloudWatch, Secrets Manager, KMS. You can make IaC (Terraform / Pulumi / CDK) decisions and live with them.
• Kubernetes in production — not just kubectl apply. You've built, broken, and fixed clusters; you know when to reach for a service mesh and when not to.
• Distributed-systems judgment — consistency models, idempotency, retries, backpressure, queueing, the failure modes of multi-region. You've debugged at least one outage that taught you something you still cite.
• Data infrastructure experience — picked and run an OLTP database at scale, built or operated an ETL/ELT pipeline, and stood up a warehouse (Snowflake / BigQuery / Redshift / ClickHouse / Iceberg-on-S3 / equivalent).
• Pragmatism about build-vs-buy. You've replaced custom orchestration with managed services and the other way around, and can explain why each was right.
• Clear writing. This role makes durable decisions; we want them legible.

Nice to have:

• Multi-tenant SaaS at scale, including noisy-neighbor and per-tenant isolation work.
• Familiarity with Effect.ts (we are heavy users) — or willingness to learn it; you don't need to love it, but you need to live with it for the parts of the codebase that stay TypeScript.
• Experience with Temporal / Airflow / Dagster / Flink / Kafka / Debezium / dbt at production scale.
• Streaming and CDC patterns; experience with Postgres logical replication.
• AI/ML infrastructure: vector stores, GPU-backed inference, embedding pipelines, prompt/response observability.
• ReBAC / Zanzibar-style authorization (we use OpenFGA).
• Background in voice / telephony / document processing pipelines.

How we work

• Small founding team — under ten people — building the system distributor sales, procurement, and dispatch teams will run on. You will be the most senior infra voice in the room.
• San Francisco — 4 days a week in the office. Foundational platform decisions go faster at a whiteboard with the team in the room.
• We ship to production frequently and trust each other to do it.
• We write specs (/specs) and architecture docs (AGENTS.md per directory) before big changes. We expect the same of you.

Compensation

• Base: $200,000 – $260,000, depending on level and experience.
• Equity: 0.5% – 1.5%. This is a founding-engineer grant; the range reflects the spread between senior and staff/principal.
• Health / dental / vision; 401(k); commuter benefits.
• Hardware + AI coding budget and a real office in SF.

Apply

Email hiring@matterhaul.com with a short note on a system you built that you're proud of and one decision in it you'd undo today. Resume optional, story not. ~ ~