Reliability & failover
High-availability architecture, failover behaviour and resource efficiency across on-prem and cloud clusters — designed so a switch is a non-event.
Senior Staff Database Reliability Engineer at Diligent. I keep large multi-engine estates fast, observable and boring — across SQL Server, PostgreSQL, MariaDB, ScyllaDB and AWS.
High-availability architecture, failover behaviour and resource efficiency across on-prem and cloud clusters — designed so a switch is a non-event.
Wait statistics, execution plans, Extended Events, pg_stat_statements. Query, index and configuration tuning that holds under real load.
Alerting that means something, and PowerShell / T-SQL / AWS CDK tooling that removes the manual step instead of documenting it.
Clustered instances used to land on a node sized for someone else's workload. I led the design of allocation that follows the workload through failover, so post-failover contention disappeared and performance stayed consistent across every node.
One interface to operate and analyse a large multi-engine estate — inventory, metrics and exports in a place engineers actually open, instead of a hundred consoles.
Flight planning for glider pilots. My own product, end to end — the same reliability instincts, in thinner air.
Own reliability, performance and operational stability of critical data platforms across relational and distributed engines, on-prem and in AWS.
Performance and availability work across a growing multi-engine platform. London.
Payments infrastructure. Cambridge.
Tuned server architecture from wait statistics, wrote the internal T-SQL standards, implemented Always On Availability Groups as the DR backbone.
Databases for stand-alone QA instruments, integrated with customer LIMS and MES systems under 21 CFR Part 11.
ERP and reporting work, plus EDI integrations across manufacturers and couriers. Where the performance obsession started.
Engineer's degree, Computer Science
Bialystok University of Technology · 2006—2011
Got a platform that has to stay up?
Reliability reviews, performance deep-dives, or a hard production problem nobody has cracked yet — start with an email.