Senior ML Engineer | Kimchi (Inference Stack Optimization)

CAST AI is hiring a Senior ML Engineer for Kimchi’s inference stack, serving large language models from its GPU fleet and customer Kubernetes clusters. The role focuses on improving throughput, latency, cold starts, and cost per token through engine tuning, storage optimization, quantization, and distributed inference.

You will own optimization decisions and technical direction, using reproducible benchmarks to guide changes. Candidates need at least five years building production ML systems, strong Python, practical experience with vLLM, SGLang, or TensorRT-LLM, and knowledge of multi-GPU infrastructure and memory behavior. Italy is among the listed hiring locations in this remote-first role. Visa sponsorship and work permits are unavailable.

Highlights

• Optimize prefix caching, scheduling, weight loading, and prefill/decode disaggregation.
• Work with Python, PyTorch, Kubernetes, and GPU inference engines.
• Benefits include equity options, learning and equipment budgets, and 10% personal development time.