技术人员，TPU 性能工程 at Inferact

部门：Research & Engineering

语言

补偿

SGD 200,000 – SGD 400,000/yr

工作方式: 现场办公
地点: 新加坡
职位类型: 全日制
级别: not_specified
发表: 2026年6月26日

说明

Inferact 正在招聘一名 TPU 性能工程师，使用 JAX、XLA、Pallas 及相关工具构建并优化 vLLM 的 TPU 后端、编译器集成、运行时路径和基准测试基础设施。该职位专注于提升 Google TPU 上生产环境模型服务的正确性、延迟和吞吐量，职责涵盖内核和推理路径优化、性能分析及基准测试。工作地点为新加坡，年薪为 200,000 至 400,000 新加坡元（S$），另加股权，并提供医疗、牙科和视力保险，以及视具体情况提供的签证赞助。

所需技能

JAX
XLA
Pallas
TPU 执行
性能分析
基准测试
机器学习内核
编译器技术

附加信息

Minimum Education: Bachelor's degree or equivalent experience in computer science, engineering, systems, machine learning, or a similar field
Required Previous Experiences: Hands-on experience building or optimizing TPU workloads using JAX, XLA, Pallas, or related compiler and runtime tooling
Experience optimizing ML kernels or inference paths, including attention, GEMM, sampling, KV cache, fused kernels, or backend runtime paths
Strong performance profiling and benchmarking skills using measurements, compiler artifacts, correctness tests, and reproducible benchmarks
Preferred Qualifications: Experience with vLLM, SGLang, TensorRT-LLM, XLA-based serving, or other LLM inference systems
Familiarity with batching, KV cache, decoding, serving tradeoffs, and backend performance constraints in production inference systems
Experience with compiler technologies such as XLA, MLIR, LLVM, Pallas, or other kernel DSLs, including lowering, fusion, and backend code generation
Knowledge of quantization methods such as INT8, FP8, mixed precision, or TPU-specific numeric formats, including accuracy and performance tradeoffs
Contribution to vLLM, JAX/XLA, Pallas, PyTorch/XLA, compiler projects, or other open-source ML infrastructure
Building TPU benchmarking infrastructure or automated performance regression detection for accelerator workloads
Work with Google TPU ecosystem stakeholders, accelerator platform teams, or early-access programs to ship backend, compiler, or inference performance improvements

福利和津贴

签证担保

关于 Inferact

Inferact

Information

Inferact is a startup founded by creators and core maintainers of vLLM, an open-source LLM inference engine. Its mission is to grow vLLM as the world's AI inference engine and accelerate AI progress by making inference cheaper and faster.

在招职位: 42
30 天内新增: 10
成立于: 2026

查看完整公司资料