WeWorkBuddy

[酷工作] [Microsoft] GPU Model Optimization 实习生 - 北京/苏州/上海

最近我们组在招 GPU Model Optimization 实习生,如果你对 LLM、GPU、CUDA、AI Infra 感兴趣,欢迎来聊~ 我们做的不是 Demo,也不是简单调 API,而是真正让大模型能够高效跑在生产环境。 平时主要会接触到: ✅ LLM / Multimodal Model Serving ✅ GPU Kernel Optimization ✅ CUDA / Triton ✅ Quantization ✅ Speculative Decoding ✅ KV Cache & Batching ✅ Distributed Training & Inference ✅ AI Agent for Model Optimization 一句话来说,就是研究如何让大模型更快、更便宜、更稳定地运行。 ⸻ 你可能会做什么 优化 LLM 和多模态模型推理性能 做 GPU Profiling,分析性能瓶颈 编写 CUDA / Triton Kernel 优化模型延迟、吞吐和 GPU 利用率 开发自动化 Benchmark 和 Performance Tool 和模型

Read original source
Source reference for [酷工作] [Microsoft] GPU Model Optimization 实习生 - 北京/苏州/上海. Screenshot not yet available.
Source record · preview pending · www.v2ex.comOpen the source for the full page.

Source record

Research analysis

Source record

Source account: 最近我们组在招 GPU Model Optimization 实习生,如果你对 LLM、GPU、CUDA、AI Infra 感兴趣,欢迎来聊~ 我们做的不是 Demo,也不是简单调 API,而是真正让大模型能够高效跑在生产环境。 平时主要会接触到: ✅ LLM / Multimodal Model Serving ✅ GPU Kernel Optimization ✅ CUDA / Triton ✅ Quantization ✅ Speculative Decoding ✅ KV Cache & Batching ✅ Distributed Training & Inference ✅ AI Agent for Model Optimization 一句话来说,就是研究如何让大模型更快、更便宜、更稳定地运行。 ⸻ 你可能会做什么 优化 LLM 和多模态模型推理性能 做 GPU Profiling,分析性能瓶颈 编写 CUDA / Triton Kernel 优化模型延迟、吞吐和 GPU 利用率 开发自动化 Benchmark 和 Performance Tool 和模型

Available context

Recent records to compare: “Google rebrands its Gemini Spark AI agent as Tasks and makes it free in six countries; in the US it can call businesses and buy items with Google Pay” (Techmeme, Nov 19, 2026); “OpenAI says agents now complete 46% of its researchers' 2-to-4-hour tasks on their own, vs. 23% in March, median researcher's agent use hit $2,400/day” (Techmeme, Nov 19, 2026). Shared themes are navigation hints; they have not been verified as the same need or independent corroboration.

Research status

This record has not been assigned a sufficiently supported main content category. No commercial or performance conclusion is generated.

Missing evidence

The main event, affected audience and supporting primary source need review.

Next step

Read the original and identify the decision it supports before choosing an analysis framework.