Backend Engineer: Scalable Rust/C++ API for AI Inference

SpaceXSI; Greater London; On-site; GBP 110,000 - 150,000; Full time

Benefits offered by this job

Equity

Healthcare and dental coverage

Cash plan

Income protection

Life insurance

Job summary

SpaceXSI is seeking a backend engineer to build and operate scalable production infrastructure for model serving. You will work primarily with Rust or C++, focusing on high-throughput inference, low latency, and reliability across distributed systems. Familiarity with Docker, Kubernetes, and gRPC will be highly beneficial.

You will own end-to-end responsibilities from API design to observability and scaling, ensuring robust deployment of billions of tokens per minute for developers worldwide.

Qualifications

• Expert knowledge of either Rust or C++.
• Designing, implementing, and maintaining reliable and horizontally scalable distributed systems.
• Knowledge of service observability and reliability best practices.
• Experience in operating commonly used databases such as PostgreSQL, Clickhouse, and MongoDB.

Responsibilities

• Build the SpaceXSI API that serves our models to developers worldwide.
• Own the end-to-end system responsible for high-throughput inference, handling billions of tokens per minute with low latency and high availability, including model serving infrastructure, request routing, SDK development, rate limiting, observability, and efficient scaling.

Skills

Rust

C++

Distributed systems

Observability

Tools

Docker

Kubernetes

gRPC

PostgreSQL

ClickHouse

MongoDB

Job description

SpaceXSI is seeking a backend engineer to build and operate scalable production infrastructure for model serving. You will work primarily with Rust or C++, focusing on high-throughput inference, low latency, and reliability across distributed systems. Familiarity with Docker, Kubernetes, and gRPC will be highly beneficial.

You will own end-to-end responsibilities from API design to observability and scaling, ensuring robust deployment of billions of tokens per minute for developers worldwide.