Machine Learning Engineer(Junior- Expert）

Our Client

• Precision-First AI Lab
• Heavy-Duty Solver Engine

What You’ll Do

• Optimize inference frameworks (e.g., vLLM, SGLang), including batching, KV cache, and scheduling
• Implement model quantization and compression (FP8 / INT8 / INT4) for production
• Profile and eliminate bottlenecks across the inference stack
• Productionize advanced techniques (e.g., speculative decoding, attention optimizations)
• Improve GPU utilization, system efficiency, and serving cost

What We’re Looking For

• Strong Python; solid systems skills (C/C++ is a plus)
• Experience with LLM inference frameworks and source-level optimization
• Good understanding of Transformer inference (prefill, decode, KV cache)
• Familiarity with GPU architecture and performance profiling
• Knowledge of quantization techniques
• 3+ years in inference optimization / HPC / related fields

Nice to Have

• CUDA / Triton and GPU kernel optimization
• Distributed inference (TP/PP) or MoE experience
• Advanced quantization (GPTQ, AWQ, SmoothQuant)
• Contributions to open-source frameworks or relevant publications