Description & Requirements
Bloomberg runs on financial data. Across datasets spanning every asset class—from alternatives and reference data to corporate actions, filings, benchmarks, point-in-time histories, and derived analytics—every record must be correct, durable, globally available, and served with low latency.
Our team builds Bloomberg’s distributed OLTP storage platform. We operate petabytes of critical financial data across more than 10,000 MySQL shards, using InnoDB and MyRocks as storage engines and Vitess for sharding and cluster management.
We build custom read and write APIs that give Bloomberg engineers safe, predictable access to financial data without exposing the complexity—or operational risk—of sharding, replication, compaction, failover, and fleet management.
This is a hands-on software engineering role at the boundary between distributed systems and database internals. You will build the software that operates and evolves this fleet, diagnose difficult production behavior at source-code depth, and contribute fixes and features upstream to Vitess and the MySQL/MyRocks/RocksDB ecosystem.
We solve recurring operational problems in software. Where practical, we contribute generally useful changes upstream rather than carrying private patches indefinitely.
We’ll trust you to:
- Design and implement capabilities that safely operate and evolve a 10,000+ shard Vitess/MySQL/MyRocks fleet across regions.
- Work in Go and C++ across Vitess, MySQL, MyRocks, and RocksDB, using Python where it is the right tool for automation, testing, and analysis.
- Improve query serving, replication, automated failover, online schema changes, VReplication and resharding, backup and restore, major-version upgrades, compaction, capacity management, and workload isolation.
- Diagnose production issues across query execution, transactions, locking, replication, storage engines, filesystems, operating systems, runtimes, and hardware.
- Build operational intelligence—including carefully tested AI-assisted tooling—that correlates metrics, logs, traces, topology, code paths, and recent changes to accelerate diagnosis and support safe, auditable remediation.
- Participate in a sustainable on-call rotation and turn recurring incidents and operational work into durable platform improvements.
- Lead technical designs and incident investigations, mentor engineers, and contribute to the open-source projects on which the platform depends.
You’ll need to have:
- Significant experience designing, building, or operating distributed databases or production infrastructure at scale.
- Deep expertise in at least one of the following areas, with an interest in working across the broader stack:
Vitess or a comparable sharded SQL system, including query routing, topology, VReplication, resharding, or cluster management;
MySQL server internals, replication, high availability, query execution, InnoDB, backup, and recovery;
MyRocks, RocksDB, or another LSM-based storage engine, including compaction, write amplification, caching, concurrency, and performance analysis.
- Strong production software engineering skills in Go and/or C++, including the ability to understand and modify large, mature systems codebases.
- Experience debugging complex behavior through profiling, tracing, thread and process analysis, source-code inspection, operating-system telemetry, and controlled reproduction.
- Experience making high-risk infrastructure changes safely through testing, staged rollout, observability, fault containment, and rollback.
- Clear technical judgment and the ability to communicate designs, trade-offs, incidents, and operational risks.
- A degree in Computer Science, Engineering, Mathematics, or a related discipline, or equivalent practical experience.
We’d love to see:
- Accepted upstream contributions to Vitess, MySQL, MyRocks, RocksDB, or another database or storage project.
- Experience operating database fleets at thousands-of-shards scale.
This role combines database internals, fleet-scale distributed systems, production ownership, and meaningful open-source impact. Join us in London to shape the storage platform behind some of Bloomberg’s most important financial data systems.