Lead Engineer Trading Platform Engineering
CoinDesk is a media outlet that covers the cryptocurrency industry.
Funding
Investors
Projects
About Coindesk
CoinDesk is a news website that focuses on cryptocurrency, blockchain technology, and digital finance. It offers analysis, data, and insights on market trends, crypto regulations, and innovations in the blockchain space. CoinDesk also hosts events, including the annual Consensus conference.
Skills
About the Role
You will design and build high performance trading systems using Java and C, optimize Linux kernel and user space components, and develop fast networking pipelines to meet deterministic latency requirements. You will work across memory management and device drivers to reduce tail latency. You will collaborate to deliver robust production systems with clear SLOs and reproducible performance tests.
Requirements
- Expert-level proficiency in Java
- Deep understanding of Linux kernel internals including memory management driver models and scheduler tuning
- Practical experience in end to end latency optimization
- Experience with socket options buffer sizing epoll io_uring RSS and NIC queue configuration
- Understanding of sequencer based total order broadcast partial order broadcast and the tradeoffs between latency throughput and availability
- Solid understanding of concurrency synchronization primitives and memory models
- Measurement driven decision making using benchmarks and profiles
- Commitment to high code quality through rigorous reviews mentoring and clear documentation
Responsibilities
- Design build and optimize high performance trading systems using Java and C
- Engineer performance critical components across user space and kernel interfaces focusing on memory scheduling I O and networking paths
- Develop and tune networking pipelines using multicast unicast UDP and TCP and implement kernel bypass or fast path networking such as DPDK or XDP
- Interact with and optimize Linux kernel subsystems including memory management schedulers device drivers and filesystems
- Optimize systems for deterministic latency through CPU isolation thread pinning NUMA locality cache awareness and reduced lock contention
- Apply distributed systems patterns such as sequencer total order broadcast for strong consistency
- Use profiling and debugging tools like perf strace eBPF and flame graphs to identify bottlenecks
- Collaborate cross functionally to deliver robust production systems with clear SLOs reproducible performance tests and runbooks