AI Machine Learning Software Engineer, up to Staff (Taipei)
Core
Architect high-performance software for AI engines and define strategy for deploying LLMs and VLMs on power-constrained hardware.
Role type
Senior/Staff IC machine learning software engineer (embedded AI)
Builds
Qualcomm AI Engine Direct (QNN), AI SW stack, and optimized neural network inference for mobile, edge, and IoT products.
Domain
Semiconductor / Embedded AI / Generative AI
Deliverable
production ML models
Required skills
C, C++, Python, neural network deployment, quantization, model compression, Transformer/LLM optimization, embedded inference frameworks (QNN, TFLite, NCNN, ONNX), performance tuning (latency, memory, power), software profiling and debugging.
Preferred skills
embedded Linux development, video/image processing, computer vision, LLM/VLM pipelines (fine-tuning, evaluation), Qualcomm AI Stack (AIMET), hardware accelerators (Hexagon DSP).
Technologies
QNN, AIMET, Hexagon DSP, TFLite, NCNN, ONNX, Transformer, ViT, CLIP.
Responsibilities
Lead development of AI SW stack framework enhancements for optimal resource usage; transition research (quantization, efficient encoders) into production-ready solutions; drive optimization algorithms for ML operators/layers specific to Generative AI; evaluate and optimize neural networks' runtime performance; develop software tools for profiling and debugging; enable state-of-the-art network models and new AI SW features for customers.
Seniority
Senior/Staff, hands-on IC with architectural leadership
