Senior Performance Engineer

Intercontinental Exchange · Atlanta

Apply on the company website

Job Purpose

Intercontinental Exchange (ICE) is hiring a Senior Performance Engineer to help own performance for our derivatives trading platform — order entry, matching engine, and market data distribution. These are latency-sensitive, business-critical systems where a regression measured in microseconds is a business event.

You will join a small, established performance engineering team. We are hiring a senior engineer to take direct ownership of a substantial share of the platform's performance work. This is not a support seat — you will run campaigns and own conclusions from day one, with an experienced team around you.

The role spans the full performance lifecycle: designing and running test campaigns, profiling pre-production code to surface bottlenecks before release, and diagnosing degradation in production. It is hands-on and deeply technical, suited to an engineer who is equally comfortable profiling a JVM under load, reading a flame graph, and explaining a latency regression to a development team in terms they can act on.

You will use AI tooling to accelerate test design, result analysis, and incident diagnosis — while remaining personally accountable for measurement quality and for every technical conclusion you present.

Responsibilities

• Own end-to-end performance validation for assigned areas of the derivatives trading platform: order entry, matching engine, and market data distribution paths.
• Design and execute load, stress, and capacity test suites against pre-production and production-like environments.
• Profile pre-production services to identify bottlenecks across CPU, memory, I/O, threading, garbage collection, network, and database layers, and partner with development teams to resolve findings before release.
• Establish and maintain production performance baselines; proactively identify degradation and emerging risk.
• Investigate latency and capacity events in production, producing formal root cause analyses and corrective action plans.
• Contribute to capacity planning and infrastructure sizing recommendations grounded in measured behavior and growth projections.
• Plan and execute non-functional testing — resiliency, failover, scalability, and capacity validation. Expect this to be roughly 20% of your time.
• Document methodology, results, and findings to a standard suitable for internal engineering audiences and, where applicable, audit and regulatory review.
• Contribute to the team's shared tooling, harnesses, and measurement practice.

AI-Assisted Performance Engineering

• AI tooling is part of the standard workflow here, not an experiment. You will use it to draft workload models and test scaffolding from production metrics, SLAs, and prior incidents; to accelerate analysis of profiler output, GC logs, traces, and packet captures; and to shorten first-pass RCA and produce engineering- and audit-ready write-ups.
• The bar is that you validate what comes back. You own the statistical soundness of the workload model, the identification of the bottleneck, the stated root cause, and the corrective actions — regardless of what produced the first draft.

Knowledge and Experience

• Bachelor's degree in Computer Science, Engineering, or a related technical discipline, or equivalent practical experience.
• 5+ years in performance engineering and testing.
• Demonstrated experience designing workload models and executing performance test campaigns against distributed, multi-tier systems, using tools such as JMeter, Gatling, k6, or LoadRunner.
• Demonstrated experience diagnosing performance problems in latency-sensitive systems using profilers and system tooling — for example perf, eBPF/bpftrace, async-profiler, JFR, flame graphs, or Intel VTune.
• A rigorous statistical approach to measurement: you understand why averages mislead, how coordinated omission distorts results, and how to build tests that produce defensible numbers.
• Excellent written and verbal communication, including the ability to explain a latency distribution to a room that contains both kernel engineers and business stakeholders.
• Deep working knowledge in at least two of the following, with working familiarity across the rest:
  • Linux internals as they relate to performance — scheduling, memory management, interrupt affinity, kernel and network tuning.
  • JVM internals — the Java memory model, JIT behavior, escape analysis, GC algorithms and their pause characteristics (G1, ZGC, Shenandoah), off-heap memory management.
  • Networking fundamentals — TCP/IP, UDP multicast, packet capture and analysis, and measurement of network-induced latency and loss.

Preferred Knowledge and Experience

• Proficiency in Java, C++, or Python.
• Applied experience using AI coding agents in a routine performance or development workflow — test-script authoring, log and metric analysis, RCA drafting. Hands-on experience with Claude Code or a comparable agent, including agent skills or MCP servers connecting agents to APM platforms, logs, or test harnesses, is a plus.
• Experience with low-latency or high-throughput systems, ideally in financial services, trading, or market data.
• Kernel-bypass and accelerated networking: Solarflare/Onload, DPDK, RDMA, or FPGA-assisted paths.
• Hardware and colocation tuning: BIOS/firmware configuration, C-state and frequency management, PTP/PPS time synchronization, hardware timestamping.
• APM platforms such as Dynatrace, AppDynamics, or OpenTelemetry.
• Chaos engineering, resiliency testing, or disaster recovery validation.

Schedule

• Onsite in our Atlanta office five days per week.
• No on-call rotation.
• Occasional weekend work to support major platform releases. This is infrequent and scheduled in advance.

#LI-JW1

More jobs at Intercontinental Exchange · Quant Development jobs in Atlanta