[ FIXED PLAN ] →
48h deployment plan. Architecture, sourcing, milestones, budget: one page. Fixed-price.
Ideal way to scope and price an engagement before signing a retainer.
> services / deploy
DevOps as a Service for crypto, AI and ZK teams
An embedded DevOps team designs, sources hardware for and deploys your Web3, LLM or ZK infrastructure end-to-end. Repo, IaC, monitoring and runbooks land as one package, not "docs later".
> what's included
Fixed scope-of-work. What we own, what stays on your side: no grey zones.
| Scope | Owned by us | Owned by you |
|---|---|---|
| Architecture design and hardware sourcing | ✓ Owned | - |
| Terraform / IaC, repository, secrets | ✓ Owned | - |
| Kubernetes cluster, network topology | ✓ Owned | - |
| Monitoring, alerts, dashboards, runbooks | ✓ Owned | - |
| Production deploy (app / validator / inference) | ✓ Owned | - |
| Business goals, SLA targets, budget | - | ✓ Owned |
| Validator keys / private key material | - (HSM/KMS, we don't custody) | ✓ Owned |
| Billing with the cloud provider | - | ✓ Owned |
> how we source hardware
We're not locked to one cloud. A mix of tier-1 providers (AWS, GCP, Azure, Hetzner, OVH, Latitude), bare-metal partners (Latitude.sh, OpenMetal, DataPacket) and regional operators in 12+ countries. For each project we pick by latency, price, supply window and jurisdiction.
For burst scenarios (incentivized testnets, hard-forks, new GPU launches) we keep open relationships with suppliers who actually have capacity on short notice. Typical supply window for 100 nodes: ≤72 hours. For H100/H200 it's 7-14 days depending on region.
If you already have a cloud account or owned hardware, we work on top of it. We don't force migrations.
> stack we ship with
Web3:
Cosmos SDK
Geth
Reth
OP Stack
Arbitrum Orbit
Polygon CDK
EigenDA
Celestia
AI / LLM:
vLLM
Triton
TensorRT-LLM
NVIDIA H100 / A100
Ray
Kubeflow
ZK:
SP1
RISC Zero
Boundless
Brevis
Jolt
Halo2
DePIN:
Filecoin
Akash
Render
io.net
Gensyn
Platform:
Kubernetes
Terraform
Ansible
Prometheus
Grafana
Loki
OpenTelemetry
PagerDuty
> engagement models
Start small. Scale up as trust builds.
[ RETAINER ] →
Monthly engagement. Full deploy + ongoing operate. On-call, patching, releases: one contract.
Best when infra is in production and you need a stable team behind it.
[ T&M ] →
Time & materials. For burst scenarios: incentivized testnets, hard-forks, migrations, emergency hire-outs.
When the volume is unknown but the urgency isn't.
> onboarding timeline
From the first call to production-ready infrastructure. Typical window: 4 weeks.
Day 0-2
Discovery + Plan
1 call. Scope, stack, regions, deadlines. One-page deployment plan in 48h with budget and milestones.
Week 1
Foundation
Hardware sourcing, cloud account setup, baseline Terraform, repository, secrets, CI/CD.
Week 2
Stack live
Kubernetes cluster, monitoring (Prometheus / Grafana / Loki), first runbook iteration, load testing.
Week 4
Handoff / Operate
Production deploy, signed SLA, hand-off into on-call. From here it's the operate phase: the pager lives with us.
> what we'd build for you
Real deliverables across the four ICPs we serve.
A validator set across 3 regions with key isolation and slashing alerts. Full package: Cosmos SDK / Geth / Reth, signing monitoring, distributed locks, HSM/KMS workflow for keys.
A vLLM inference cluster on A100/H100 with autoscaling and request tracing. GPU scheduler, cost-per-token dashboards, p95 latency SLO, model-rollback playbooks.
A prover farm with GPU scheduler and proof-job retry queues. SP1 / RISC Zero / Boundless / Brevis, per-circuit benchmarks, deadline-aware job prioritization.
A node-onboarding pipeline turning a fresh VM into a working network participant in <10 min. Filecoin / Akash / io.net / Render / Gensyn, payout reconciliation, regional rebalancing.
> SLA tiers
Three coverage levels after handoff into operate. Pick by criticality.
| Tier | Response p95 (Sev-1) | Coverage | Incident report | Engineer hours / mo |
|---|---|---|---|---|
| Bronze | 30 min | Business hours, 5×8 | Within 48h | 40 |
| Silver | 15 min | 24/7 on-call rotation | Within 24h | 80 |
| Gold | 5 min | 24/7 with dedicated engineer | Within 12h | 160+ |
> related services
> FAQ
A single L1/L2 validator: 5-10 business days. GPU inference cluster across 3 regions: 10-14 days. Burst up to 100 nodes for an incentivized testnet: 72h. A full multi-region platform: 4 weeks.
Yes. That's the most common setup: provider billing stays with you, we come in as a DevOps team on top. If needed, we source hardware via our own supplier contracts.
A Git repository with Terraform/IaC, architecture documentation, secrets in KMS, monitoring (Prometheus / Grafana / Loki) with dashboards, versioned runbooks, production deploy, and a hand-off document. Everything lives in your accounts. We don't hold anything hostage.
You do. HSM/KMS workflow where keys never leave your control. We sign via process, we don't custody material.
Fixed-price for 48h. Amount depends on scope: send a brief, we'll respond within 24h. The plan stands on its own even if you don't continue: you walk away with architecture and a budget estimate on paper.