Performance & Modeling Engineer
Teknik, data och digitalt · Mjukvaru- och webbutveckling · Mjukvaruutveckling · Maskininlärning · Datavetenskap
I korthet
Qualcomm Cloud AI is looking for a Performance & Modeling Engineer in Beijing, China, to optimize state-of-the-art AI models on their platforms. This role involves deploying and analyzing AI models, driving performance optimizations, and supporting customer deployments.
Ansvarsområden
- Deploy and optimize LLM, MoE, and multimodal models on Qualcomm Cloud AI platforms.
- Analyze and improve model accuracy, latency, throughput, memory utilization, and scalability.
- Drive graph-level, framework-level, and system-level optimizations.
- Benchmark AI workloads and identify performance bottlenecks across the software and hardware stack.
- Work closely with compiler, runtime, kernel, and architecture teams to improve model enablement and performance.
- Support customer deployment, bring-up, performance tuning, and production readiness.
Krav
- Master's degree or above in Computer Science, Artificial Intelligence, Mathematics, Electrical Engineering, or related fields.
- Good verbal and written English communication skills.
- Strong experience with Python, C++, and Linux development environments.
- 5+ years of experience in Cloud AI, LLM deployment, or AI infrastructure.
- Deep understanding of Transformer architectures, LLMs, and MoE models.
- Strong experience in graph optimization and model compilation workflows.
- Strong experience in model accuracy and performance optimization.
- Strong experience with one or more AI frameworks, including PyTorch, ONNX, TensorFlow, or JAX.
- Experience with AI serving frameworks such as vLLM, SGLang, TGI, TensorRT-LLM, or LMDeploy.
Önskade kvalifikationer
- Familiarity with AI accelerator platforms and large-scale inference systems.
- Experience with Quantization, KV-Cache optimization, speculative decoding, continuous batching, Prefix Caching, Parallelism Strategies and distributed inference.
- Experience with performance modeling and analysis methodologies, including Roofline analysis, bottleneck identification across software and hardware stacks.
- Familiarity with open-source inference frameworks such as llama.cpp, Ollama, or MLC-LLM.
- Experience with Kubernetes, Docker, and cloud-native AI infrastructure.
Förmåner
- Qualcomm is an equal opportunity employer.
- Qualcomm is committed to providing an accessible process for individuals with disabilities.
- Qualcomm is committed to making its workplace accessible for individuals with disabilities.
Färdigheter
PythonC++LinuxPyTorchONNXTensorFlowJAXvLLMSGLangTGITensorRT-LLMLMDeployKubernetesDocker
#AI#LLM#Cloud AI#Performance#Modeling#Optimization#Inference
Om Qualcomm
Enabling a world where everyone and everything can be intelligently connected.
Enabling a world where everyone and everything can be intelligently connected.\n\n
Företag
QualcommPublicerade jobb
för 1 vecka sedan
Anställningstyp
Heltid
Arbetsform
På plats
Erfarenhetsnivå
Senior
Platser
Beijing, China
Kvalifikation
Masterexamen, Kandidatexamen, Doktorand
Sökande
Ansök tidigt
Liknande jobb
Ericsson
Budapest, HungaryStudent Data Engineer
Heltid Ansök tidigt Publicerad för 4 dagar sedan
Alstom
Charleroi, BelgiumSoftware Designer
Heltid Ansök tidigt Publicerad för 4 dagar sedan
Sandvik
Tampere, FinlandSenior Electrical Engineer
Heltid Ansök tidigt Publicerad för 4 dagar sedan
Accenture
Copenhagen, DenmarkUtilities Strategy Manager
Heltid Ansök tidigt Publicerad för 4 dagar sedan
Ericsson
Łódź, PolandJunior Software Developer
Heltid Ansök tidigt Publicerad för 4 dagar sedan
Scania Group
Koblenz, GermanySenior Full-Stack Anwendungsentwickler
Heltid Ansök tidigt Publicerad för 3 dagar sedan