[Skip to main content](#main-content)

[OpenNash Jobs](https://jobs.opennash.com/)
[Hiring report](/reports/hiring/) · [All Xiaomi jobs](https://jobs.opennash.com/companies/xiaomi/) · [Search jobs](/search/)

Xiaomi · 深圳

# 高级基带工程师

深圳 

Listed on Xiaomi ’s official careers site. Posted August 6, 2026. Check that it is still open and apply there.
[Apply on Xiaomi ’s careers site →](https://xiaomi.jobs.f.mioffice.cn/index/position/7670813273545853226/detail)

[OpenNash Jobs](https://jobs.opennash.com/) / [Xiaomi ](https://jobs.opennash.com/companies/xiaomi/) / 高级基带工程师

## Job description
- 负责AI加速芯片技术方案架构设计和性能优化，包括基于3D封装的NPU+DRAM的方案以及PIM和HBF等方案；
- 负责芯片微架构评估与选型，输出技术可行性分析和Benchmark报告，与算法团队协同，完成模型量化（INT4/INT8）、稀疏化等部署适配；
- 负责AI加速芯片与手机端侧大模型团队进行端侧模型适配调优，算力能效提升等核心技术攻坚；设计面向大模型推理的硬件加速方案，包括算子映射、数据流优化、内存带宽管理；
- 负责大模型在手机端侧的推理性能优化，实现低延迟、低功耗部署，针对特定硬件平台（骁龙/天玑/自研芯片）进行算子级深度优化；
- 跟踪业界AI芯片技术演进（高通HTP、联发科APU、自研NPU），制定技术路线图；构建端侧推理性能评测体系，输出性能调优报告；

## Qualifications

- 硕士及以上学历，计算机体系结构/微电子/集成电路相关专业
- 5年以上AI芯片或SoC设计经验，熟悉NPU/GPU/DSP微架构
- 精通至少一种硬件描述语言（Verilog/SystemVerilog）或AI编译器框架（TVM/MLIR）
- 熟悉主流端侧AI推理框架（SNPE、QNN、MNN、TFLite）
- 有Transformer/LLM模型硬件加速经验者优先
- 有流片经验或芯片公司背景优先

## At a glance
- Posted by the employer: Aug 6, 2026
- First seen by OpenNash: Oct 19, 2026
- Last checked: Oct 19, 2026
- Listed for: 74 days
- Location: 深圳[Apply on Xiaomi ’s careers site →](https://xiaomi.jobs.f.mioffice.cn/index/position/7670813273545853226/detail)

## More jobs at Xiaomi 

### [AI Agent开发工程师](https://jobs.opennash.com/companies/xiaomi/jobs/b6d78153a2ad67ccdb4f/)

深圳

### [BSP底层专家](https://jobs.opennash.com/companies/xiaomi/jobs/d5d898b5e9a64be314bb/)

深圳

### [GTM专员](https://jobs.opennash.com/companies/xiaomi/jobs/e202fdb1f52d0445d09a/)

深圳

### [Sram版图设计工程师](https://jobs.opennash.com/companies/xiaomi/jobs/25cc7ad2bad44f09c0c0/)

深圳

### [SoC芯片验证实习生-2027届](https://jobs.opennash.com/companies/xiaomi/jobs/d506fd60a52b9a6ffd60/)

深圳

OpenNash lists jobs from employers’ own careers sites and does not handle applications.
Dates come from the employer where stated; “first seen” and “last checked” are our own observations.

OpenNash Jobs · Official listings with dates and links to the employer. [Weekly hiring report](/reports/hiring/) · [Methodology](/methodology/)