Bloomberg runs on financial data. Every price, corporate action, benchmark, security master update, point-in-time view, and derived fact must be correct, fast, durable, and available at global scale.

Our team builds the distributed OLTP storage layer behind some of Bloomberg’s most critical financial data systems. We provide deterministic, purpose-built APIs that let Bloomberg engineers read and write financial data safely, while our platform absorbs the complexity of Vitess, MySQL, MyRocks, RocksDB, sharding, replication, compaction, observability, capacity management, and failure recovery.

This is a role for engineers who want to work close to the metal of distributed data systems. You will reason about query routing, storage-engine behavior, schema evolution, read-your-own-write consistency, hot shards, p99 latency, saturation, compaction debt, replication lag, backfills, and safe production rollout.

When Vitess, MySQL, MyRocks, RocksDB, or related tooling need to change to meet Bloomberg’s requirements, we are prepared to make those changes — and contribute them upstream.

Our Team

We build the store services that make Bloomberg financial data discoverable, queryable, point-in-time accurate, and reliable across the company.

Our systems expose structured, schema-evolving datasets through unified APIs while preserving consistency, low latency, operational safety, and auditability. They serve both latency-sensitive OLTP traffic and high-throughput readers that power downstream systems, including Bloomberg’s S3-based BBDS lakehouse.

This role is not about Spark internals. It is about making a distributed OLTP storage platform remain correct and performant under very different workload shapes: small deterministic reads, write-heavy dataflows, backfills, table materialization, snapshot reads, and massively parallel batch access.

We care deeply about production behavior. That means source-level debugging, disciplined benchmarking, workload modeling, high-cardinality observability, automated operations, safe rollout, and tooling that helps engineers understand exactly why the system behaved the way it did.

What’s in it for you

• Build OLTP store services supporting billions of mutations and transactions across thousands of shards.

• Work deeply with Vitess, MySQL, MyRocks, RocksDB, and the operational realities of running them at scale.

• Design deterministic, low-latency access patterns instead of relying on arbitrary SQL execution.

• Build systems for time series, bi-temporal data, point-in-time queries, versioned facts, snapshot reads, and derived table materialization.

• Diagnose and improve performance across the stack: application code, JVM, Linux, networking, MySQL, MyRocks, RocksDB, and Vitess.

• Build observability for fast production debugging using structured events, traces, high-cardinality dimensions, SLIs, and storage-engine telemetry.

• Make store services robust under both client-facing OLTP traffic and highly parallel batch-reader workloads.

• Influence the architecture of Bloomberg’s financial data platform.

• Contribute improvements back to upstream open-source systems when Bloomberg’s requirements push beyond what exists today.

You’ll need to have

• 4+ years of professional software engineering experience.

• Strong programming skills in one or more of Java, Go, C, or C++.

• Deep experience or strong interest in database internals, distributed systems, storage engines, or large-scale data infrastructure.

• Strong debugging skills across application code, infrastructure, networking, and Linux.

• Experience reasoning from first principles about throughput, latency, contention, queuing, caching, compaction, replication, saturation, and failure modes.

• A degree in Computer Science, Engineering, Mathematics, or equivalent practical experience.

We’d love to see

• Contributions to infrastructure projects such as Vitess, MySQL, RocksDB, MyRocks, or similar systems.

• Experience operating or extending sharded MySQL, Vitess, distributed SQL, or large-scale OLTP platforms.

• Experience profiling and tuning production systems with JFR, async-profiler, perf, eBPF, flame graphs, heap analysis, or custom benchmarking frameworks.

• A track record of building reliable systems where correctness, latency, and operability matter.

• A passion for mentoring engineers and raising the technical bar of a platform team.