大模型推理架构师
Tasks
- Automate tuning and resource scheduling
- Build AI compilation and model optimization pipeline
- Build AI training and inference infrastructure
- Design end to end AI system architecture
- Develop high-performance inference engine
- Develop performance analysis and observability tools
- Diagnose system failures and bottlenecks
- Implement model compiler and graph optimizations
- Improve GPU and NPU utilization
- Optimize distributed training performance
- Optimize heterogeneous compute kernels
- Optimize large scale distributed training and inference
- Reduce latency and increase throughput
- Validate and deploy new LLM system techniques in production
Perks/Benefits
- N/A
Skills/Tech-stack
Ascend C | C plus plus | C# | CUDA | CUDA kernel | Checkpointing | Communication overlap | Cutlass | Data-parallel | DeepSpeed | Distributed Data Parallel | Distributed data | FSDP | Graph optimization | Infiniband | KV cache | LLVM | MLIR | Megatron | Mixture of Experts | Model compilation | NCCL | NVSHMEM | Offloading | Pipeline parallelism | PyTorch | Python | Quantization | RDMA | RoCE | SGLang | TVM | Tensor Parallelism | TensorRT-LLM | Torch Dynamo | TorchInductor | Transformer | Triton | Triton Compiler | VLLM | Verbs | XLA | Zero
Education
Bachelor of Engineering | Bachelor of Science | Master of Science | PhD
Related jobs
-
Entry-level Full Time广州、上海3h ago
-
Entry-level Internship苏州4h ago
-
AI Foundation Model方向实习生 CNY 25K-37KC++ | Caffe | Computer Vision | Data Structures | Data Structures and AlgorithmsEntry-level Internship北京、上海4h ago
-
Ai产品经理 CNY 144K-240KArtificial Intelligence | Data analytics | Data cleaning | Data collection | Data synchronizationEntry-level Full Time深圳4h ago
-
Mid-level Full Time深圳 R4h ago
-
【27届校招提前批】算法工程师 CNY 180K-360KAgent collaboration | Asynchronous programming | Caching | Concurrency | Context engineeringNone Full Time上海4h ago
-
Entry-level Full Time北京4h ago
-
Entry-level InternshipBeijing,Beijing,China12h ago
-
Automation | Cost Optimization | Evaluation | Guardrails | JavaScriptFast hiring process | Remote workMid-level Full TimeChina20h ago
-
APIs | Agents | Cost Optimization | Data integration | EvaluationRemote workMid-level Full TimeChina20h ago
-
Mid-level Full Time东莞21h ago
-
Dba 数据库工程师-2027届 CNY 180K-300KDatabase high availability | Database performance | Database performance tuning | Go | High AvailabilityNone Full Time北京、上海22h ago
-
C# | C++ | Camera Calibration | Computer Vision | Depth EstimationMid-level Full TimeShanghai, Shanghai, China22h ago
-
Mid-level Full Time北京 R23h ago
-
大模型算法研究员-MiMo CNY 500K-500KActive Learning | C++ | Curriculum learning | Deep learning | Instruction TuningEntry-level Full Time北京23h ago
-
Entry-level Full Time北京 R1d ago
-
高级算法工程师（Nlp方向） CNY 240K-480KAgent memory | CUDA | Chroma | Client-Server | Client-Server ArchitectureSenior-level Full Time北京1d ago
-
Nlp / Llm 应用工程师 CNY 180K-360KAnthropic API | Chroma | Embedding | Genspark | Genspark RAG ArchitectureEntry-level Full Time北京 R1d ago
-
机器人VLA算法研究员 - XiaomiRobotics CNY 500K-500KDeep learning | Diffusion | Machine Learning | Mixture of Experts | Multimodal LearningMid-level Full Time北京1d ago
-
Mid-level Full Time北京 R1d ago
-
具身世界模型推理INFRA工程师 - XiaomiRobotics CNY 144K-240KDiffusion Models | Expert parallelism | FP8 | Machine Learning | Model InferenceSenior-level Full Time北京1d ago
-
Entry-level Full Time北京 R1d ago
-
Mid-level Full Time北京1d ago
-
资深机器视觉专家 CNY 180K-300KAction Recognition | Computer Vision | Deep learning | Docker | Image ProcessingMid-level Full Time北京、上海1d ago
-
Entry-level Full Time北京1d ago