The company is a small, remote-only team whose software runs production Kubernetes infrastructure around the world.
This is a Rust data plane role building the high-performance proxy and traffic-handling code behind the company's service mesh and its enterprise edition.
The code sits directly in the live request path. The work is building networking product software, not deploying or running a service mesh.
The company competes against rivals with far larger budgets and headcounts, so each engineer's output matters.
No prior experience with the company's mesh is needed, but you need a track record writing low-level production networking code.
Working hours: Americas-based candidates need at least 5 hours of overlap with Pacific Time. Central Europe-based candidates need to be reachable from 7 AM to 10 AM Pacific Time.
Responsibilities
Build, test and maintain fast Rust networking code for the data plane, including the proxy and related traffic components.
Work on live L4 and L7 traffic across TCP, HTTP/1.x, HTTP/2 and gRPC, and diagnose connection and protocol behaviour.
Add proxy features such as connection management, load balancing, retries, timeouts, backpressure, routing and graceful handling of failures.
Improve latency, throughput, and CPU and memory use through benchmarks, profiling, production telemetry and structured analysis.
Track down hard Linux, networking, concurrency and distributed systems bugs, including failures that appear in production.
Design secure service-to-service communication and reliability, working closely with control plane engineers.
Write production code, take part in design reviews, write technical documentation, and contribute to relevant open-source work on the mesh.
Requirements
You have 7+ years of professional experience building production systems software, networking products, distributed infrastructure or other performance-sensitive back-end systems.
You have personally designed, built and debugged software that processes, forwards, proxies or routes network traffic. Operating such software does not count.
You understand TCP/IP, connections, sockets and HTTP well enough to troubleshoot live traffic and protocol behaviour.
You have dealt with throughput, latency, concurrency, memory and CPU efficiency, and failures in production software.
You have strong systems programming skills in Rust, or deep production C or C++ networking experience and a clear ability to become productive in Rust. Production Rust is strongly preferred.
You have built and debugged interconnected production services, including partial failures, timeouts and resource limits.
You have taken substantial features from design and build through testing, production debugging and upkeep.
You communicate clearly in writing and speaking and can work independently on a remote distributed team.
You take ownership, reason through unclear problems, write clear designs, collaborate without ego, and ship reliable software.
Americas-based candidates must keep at least 5 hours of daily overlap with 8:00 AM to 5:00 PM Pacific Time, Monday to Friday. Central European candidates must be able to collaborate between 7 AM and 10 AM Pacific Time.
This is a production network and data plane software role, not a DevOps, SRE, platform operations, Kubernetes administration or general application backend role.
Not enough on its own: deploying, configuring or troubleshooting service meshes, Envoy, Istio, NGINX, API gateways or load balancers without building their traffic-processing code.
Not enough on its own: managing Kubernetes clusters, mesh installs, Helm charts, Terraform, CI/CD, production infrastructure or observability dashboards.
Not enough on its own: REST APIs, CRUD microservices, web applications or business logic without significant networking or systems-level work.
Not enough on its own: Rust as a hobby or course language, without strong production Rust or deep professional C/C++ systems expertise.
Not enough on its own: general Go backend work, without ownership of traffic-path, proxy, packet-processing or performance-sensitive systems software.
Preferred: Production Rust, especially asynchronous networking with Tokio and related libraries.
Preferred: Hands-on development of L4 or L7 proxies, load balancers, reverse proxies, gateways, packet-processing systems or service mesh data planes.
Preferred: In-depth knowledge of HTTP/2, gRPC, streaming, TLS and mTLS, connection pooling, flow control, retry policies and backpressure.
Preferred: Linux networking internals, kernel and user-space networking, eBPF, XDP or DPDK.
Preferred: Profiling production systems, benchmarking, reducing tail latency, and finding race conditions, deadlocks and resource contention.
Preferred: Knowledge of the company's mesh, Envoy, Istio, Kubernetes networking, service discovery and cloud-native infrastructure.
Preferred: Open-source contributions, security-minded engineering and experience reviewing complex systems code.
Conditions
Fully remote, open to candidates in the United States, Canada, South America or Central Europe on CET or CEST time.
Americas-based candidates must overlap with Pacific Time for at least 5 hours each day, between 9:00 AM and 5:00 PM, Monday to Friday. Central Europe-based candidates work an agreed CET or CEST schedule that overlaps regularly with the US engineering team.
US and Canadian applicants must already be authorized to work in that country, now and in the future, without the employer's immigration sponsorship. The company is not offering sponsorship for this role.
Candidates in South America or Central Europe may be considered under company-approved employment or independent contractor arrangements, subject to country-specific eligibility and approval.