ClickHouse Engineer

ClickHouse Engineer Singapore (Office-based) or Remote (UK-based) | Full-Time About Us Meridian Softworks is the 2026 OTC Trading Platform of the Year (). Our network handles over $50 billion in daily trading volumes across FX, Equities, and Cryptocurrency, connecting 40+ of the world's leading liquidity providers. We build trading systems that operate at the edge of what's technically possible — where nanoseconds matter and a dropped message costs real money. Our engineering team is small, senior, and deeply invested in the craft of building cutting edge, reliable, high-performance systems. The Role We're looking for a ClickHouse Engineer to share ownership of our data plant: a multi-datacentre, multi-AZ ClickHouse cluster ingesting billions of rows a day, underpinning compliance, billing, and a growing family of real-time and historical analytics. This is a core-platform engineering role, not a DBA role. You'll work across the full engineering scope of a serious ClickHouse estate: cluster operations and upgrades, schema and sorting-key design, materialised views, ingestion performance, query optimisation, workload isolation, capacity planning, storage tiering, backup and recovery, monitoring and alerting, and production support. You'll also work on the data pipelines around the plant — our capture services are Go, and our analytics foundation is Python. You don't need years of ClickHouse specifically — deep experience with another columnar or large-scale time-series database qualifies you, and a strong columnar engineer learns our estate quickly. What we do need is genuine operational depth: you have run a production data platform, owned its capacity plan and its recovery procedures, and stayed calm when it mattered. This hire completes a follow-the-sun team — based in Singapore (office-based) or London — so the platform is supported across time zones rather than by heroics. On-call is a shared, contracted responsibility, planned and paid — not goodwill. What You'll Work On Cluster operations at scale — running, upgrading and evolving a multi-AZ ClickHouse estate on Kubernetes, with rehearsed backup and recovery Schema and query engineering — table and sorting-key design, partitioning, materialised views, and query optimisation against multi-terabyte datasets Capacity and observability — a measured capacity model, storage tiering and retention, and the Grafana monitoring and alerting that keeps the plant's health visible Ingestion and data quality — correctness and completeness gates on the pipelines feeding the plant; reconciliation that proves nothing was dropped Workload isolation — keeping ingest, operations, reporting and ad-hoc analytics from treading on each other as read load grows AI-assisted operations — modern tooling (including MCP-based AI access to the estate) that lets a small team run a large platform well What We Need Technical: Columnar/OLAP database engineering — ClickHouse strongly preferred; deep experience with another columnar or large-scale time-series store (BigQuery, Redshift, Druid, kdb+ or similar) also works SQL depth — execution plans, query optimisation, and schema design for very large datasets Operational/SRE experience — you have carried production responsibility for a data platform: monitoring, capacity, incidents, recovery Linux and Kubernetes — comfort with the operational layer beneath the database A programming language — one of Python, Go, R, or MATLAB (Python preferred) Engineering discipline — you write proposals, strategies, and architectural documents and diagrams well, and you manage change properly Git — excellent version-control practice Distributed-systems fundamentals — replication, consistency, failure modes Financial services experience is a plus but not essential — domain knowledge can be learned; operational instinct cannot How you work: You own production. Calm in an incident, rigorous in a post-mortem, honest about what nearly went wrong You measure first. Topology, storage and optimisation decisions follow observed load and cost — not fashion You write things down. Runbooks, schema documentation and clean handovers are part of the craft You share knowledge deliberately. The goal is a platform supportable by a team, not a specialist You communicate across time zones. Follow-the-sun only works with clear, proactive handovers You embrace AI as a tool. You use it to amplify your own effectiveness and help shape how it operates safely around production data What You Get A serious estate — genuine scale, genuine criticality: the system of record for a live trading network, not a reporting sidecar Real ownership — shared responsibility for a business-critical core platform from day one, working directly with the platform's lead engineer An excellent team — senior engineers who care about craft, collaborate openly, and hold each other to high standards Sustainable operations — follow-the-sun coverage by design; on-call is shared, contracted and planned, not he