Staff / Principal Machine Learning Engineer, Inference - Switzerland
Inworld · Remote · Switzerland ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Staff/Principal level
- Fully remote (Switzerland)
- C++ / CUDA / Rust
- LLM Inference Optimization
- No visa sponsorship
Join the Inference team at Inworld, where you will build the infrastructure behind modular APIs serving over 10 trillion LLM tokens monthly. You will tackle ambiguous challenges in real-time multimodal inference, optimizing systems for sub-second latency and massive scale across diverse consumer sectors.
What you'll do
- Take models from research to production, handling containerization and serving optimization
- Profile code and optimize performance on NVIDIA GPUs for low latency and high throughput
- Design benchmarks and prototypes to solve unclear problems in distributed systems
- Ensure reliability and stability of inference infrastructure under high concurrent load
- Question architectural decisions to improve core latency and throughput metrics
What you'll need
- Deep understanding of serving frameworks like vLLM or TRT-LLM
- Proficiency in C++, CUDA, Rust, or highly optimized Python
- Experience with quantization, distillation, caching, and speculative decoding
- Knowledge of Kubernetes, Ray, and multi-GPU/multi-node scaling strategies
- Professional fluency in written and spoken English
- Legal right to work in Switzerland is mandatory
Nice to have
- PhD in Computer Science, Physics, Math, or equivalent practical experience
- Open-source contributions to major inference engines
- Public technical write-ups or non-trivial systems programming projects
- Experience building backend or ML systems in ambiguous environments
Stack
- vLLM
- TRT-LLM
- C++
- CUDA
- Rust
- Python
- Kubernetes
- Ray
- NVIDIA GPUs
How you'll work
- Fully remote within Switzerland
- Full-time permanent employment via Employer of Record
- No visa sponsorship available for this role
- Collaboration with US-based leadership and engineering teams