Role overview

Analog builds industrial intelligence infrastructure for environments where cloud services are not available or appropriate—air-gapped sites, regulated facilities, and physically demanding locations without internet connectivity. As a Distributed Systems Architect, you will design the bare-metal clustering, replication, and failover systems that protect clients from data loss at petabyte scale, ensuring their on-premises infrastructure survives hardware failures and node outages without external support.

Key responsibilities

Map Analog’s cloud-based components to their open-source on-premises equivalents, documenting trade-offs for each architectural choice including Kafka or Redpanda, EMQX, ClickHouse or Druid, and Ceph or MinIO.
Design multi-node clustering topologies, rack-aware replication strategies, and quorum configurations for bare-metal Kafka or Redpanda and Postgres clusters using Patroni.
Optimize Linux systems at kernel level—including parameter tuning, NUMA bindings, NVMe I/O scheduling, CPU affinity, and IRQ balancing—to eliminate hardware bottlenecks under petabyte-scale write workloads.
Produce implementation-ready blueprints and runbooks that the DevOps team can execute directly using Terraform and Ansible without requiring architectural interpretation.
Review client hardware specifications before deployment and provide clear technical approval or rejection decisions based on infrastructure requirements.
Build a reusable library of reference architectures that standardizes future air-gapped deployments across the organization.

Requirements

Minimum 10 years of hands-on experience architecting and operating distributed infrastructure systems at terabyte to petabyte scale in production environments.
Expert knowledge of Kafka or Redpanda clustering operations including partitioning strategies, in-sync replica tuning, consumer group management, log compaction, and the full operational lifecycle.
Deep expertise in Linux systems engineering covering kernel networking, TCP buffer tuning, interrupt coalescing, storage fabrics, and NUMA-aware process binding.
Proven track record designing and running highly available database clusters without managed cloud load balancers, using tools such as Patroni or Pacemaker in production.
Experience deploying and operating Ceph or MinIO on physical hardware, including CRUSH map configuration, erasure coding strategies, and performance optimization.
Ability to write clear blueprints and runbooks that other teams can execute independently without requiring your presence to interpret architectural decisions.

What makes this role worth considering

Your designs have direct, immediate consequences: there is no abstraction layer between your architecture and what ships to clients. You will be the technical authority whose decisions determine whether a client’s system survives critical failures. The work you produce compounds over time as reference architectures become the foundation for every future deployment, multiplying your impact across the organization.