From idea to production, and running long after launch.
Polaz turns a product idea into a system that serves real users and survives real failures. Architecture, cloud and Kubernetes platforms, high availability and disaster recovery across regions, AI infrastructure, networking, identity, security and the operations that keep it all up. One senior team for the whole path, so nothing falls between vendors.
What we do
Everything between "we have an idea" and "it runs in production, survives an outage and we sleep at night". We work inside your environment and leave it documented and automated.
Architecture & technical leadership
From a blank page to a design that survives growth. System boundaries, data model, build-or-buy, security model, availability targets and a delivery plan. Fractional CTO when you need one.
High availability, DR & geo-clusters
Systems that keep serving when a server, a data centre or a whole region goes down. Active-active geo-clusters, multi-cluster Kubernetes, quorum-based failover, replicated databases and disaster recovery with agreed RPO and RTO, rehearsed rather than hoped for.
Cloud & Kubernetes platforms
Private, public and hybrid cloud. vSphere, NSX and Proxmox estates, AWS, Azure and GCP, Talos and k3s clusters with Cilium BGP and Cluster Mesh, GitOps delivery and backups that are actually restored.
AI & data platforms
AI that runs on your terms. Private LLM inference and GPU capacity, retrieval (RAG) over graph, vector and full-text search, AI agents wired into your tools over MCP, and the data pipelines that feed them.
Zero-trust networking
WireGuard and NetBird overlays with group-based access, Envoy edges with health-checked failover, BGP with your upstreams, and firewalls that follow the design.
Identity & access
Single sign-on with OIDC and Kerberos, FreeIPA directories, passkeys and WebAuthn. One identity across VPN, applications and servers.
Internet edge & traffic
The layer every user touches first. Global load balancing and anycast, edge proxies with TLS automation, signed DNS across independent providers, and email infrastructure your customers can trust.
Product & backend engineering
APIs, data stores and the services behind your product, built to be operated. Observable, tested, deployed by CI and fast where it matters.
Operations & security
Monitoring and alerting, centralized logs, SIEM and vulnerability scanning, incident response, upgrades without downtime. Findings come with fixes.
Open source we build
Part of our engineering is public. Databases, identity, security protocols and developer tooling, maintained by the same people who would work on your systems.
CoordiNode
Graph-native hybrid retrieval engine for AI. Graph, vector and full-text in one transactional engine.
StructuredID
Self-hosted identity provider with OPAQUE (RFC 9807), WebAuthn, OAuth2 and OIDC.
gitlab-mcp
GitLab for AI assistants over MCP. Issues, merge requests, pipelines and code search.
structured-proxy
Config-driven gRPC to REST transcoding proxy for any gRPC service.
krb5-rs
Kerberos V5 with GSSAPI, SPNEGO and PKINIT, no system krb5 dependency.
xml-sec
XMLDSig, XMLEnc and C14N, a drop-in replacement for libxmlsec1.
coordinode-lsm-tree
Embedded storage engine with MVCC, encryption at rest and self-healing per-block ECC.
structured-zstd
Zstandard compression with streaming and dictionaries, for servers, embedded and the browser.
vue-privacy
Privacy-first consent and analytics for Vue, Nuxt and Quasar with Google Consent Mode v2.
Ways to work with us
Small team, senior people, no account managers. Pick the shape that fits; we can change it as the work changes.
Part-time retainer
A senior engineer embedded in your team for a fixed number of hours each month.
- From 40 hours per month
- Your tracker, your repositories
- Monthly report of what changed
Fixed-scope project
A defined outcome. An MVP taken to production, a platform build-out, a migration, an SSO rollout.
- Written scope and acceptance criteria
- Milestone-based invoices
- Handover with runbooks
Audit & advisory
A focused review of architecture, production readiness, resilience and disaster recovery, or security posture, before an investment round, a launch or an incident does it for you.
- One to two weeks
- Prioritised findings with fixes
- Follow-up call included
How an engagement runs
Predictable, written down, and measured.
Intro call
We learn what you run, what hurts and what success looks like.
Proposal
Scope, approach, timeline and the engagement model, in writing.
Delivery
Work in your environment, automated and reviewed, with regular updates.
Handover
Documentation and runbooks, so your team owns the result.
Tell us what you are building
A short description is enough. We reply within one business day.