Similar Jobs
See allThe challenge:
- Move throughput, latency, cold starts, and cost per token in the right direction.
- Own the inference stack's optimization, from kernels to serving configuration.
The stack:
- Python, SGLang, vLLM, TensorRT-LLM, PyTorch, CUDA-adjacent tooling, Kubernetes.
- Shared filesystems, ClickHouse, OpenTelemetry, Prometheus, Grafana, GitLab CI.
The team:
- Remote-first global environment with a fast-paced workflow.
- Equity, learning budget, hackathon, and 10% time for personal projects.
Kimchi
Kimchi is an LLM inference platform serving open models like DeepSeek and GLM from its own GPU fleet via one API line. It's made by CAST AI, a globally distributed team across 34 countries that hit unicorn status in 2026.