AI Machine Learning Software Engineer Taipei

Qualcomm Machine Learning Engineer

As a leading technology innovator, Qualcomm pushes the boundaries of what's possible to enable next-generation experiences and drives digital transformation to help create a smarter, connected future for all. As a Qualcomm Machine Learning Engineer, you will create and implement machine learning techniques, frameworks, and tools that enable the efficient discovery and utilization of state-of-the-art machine learning solutions over a broad set of technology verticals or designs.

Qualcomm Engineers collaborate with cross-functional teams to enhance the world of mobile, edge, auto, and IoT products through machine learning hardware and software.

We are looking for a Senior or Staff level Engineer to work on bleeding-edge AI technology. You will architect high-performance software for AI engines, including Qualcomm AI Engine Direct (QNN), and define the strategy for deploying Large Language Models (LLM) and Vision-Language Models (VLM) on strictly power-constrained hardware. You will collaborate with cross-functional teams (HW/SW architecture) to enhance the world of mobile, edge, and IoT products.

This is a great opportunity to innovate and develop leading-edge products around best-in-class Qualcomm AIoT devices.

• Architectural Leadership:
  Lead the development of AI SW stack framework enhancements for optimal resource usage while running complex Transformer-based networks (LLM, ViT) on Qualcomm hardware.
• Research to Production:
  Lead efforts in transitioning research (e.g., new quantization techniques, efficient CLIP encoders) into production-ready solutions, enabling real-world applications and commercial impact.
• GenAI Optimization:
  Drive the development of optimization algorithms for ML operators/layers specific to Generative AI (e.g., KV-cache optimization, attention acceleration) within the Qualcomm AI SW stack.
• Performance Tuning:
  Evaluate and optimize neural networks' runtime performance (latency, memory, power) and accuracy using tools like AIMET and QNN SDK.
• Software Development:
  Develop software tools for profiling and debugging to support the rapid deployment of new neural networks.
• Feature Enablement:
  Work with customer teams to enable state-of-the-art network models and new AI SW features to meet customer use-cases, and collaborate with AI hardware teams to continuously improve our AI solution.

• Master's degree in Electrical Engineering, Computer Science, Mathematics, Physics, or a closely related field with 5+ years of relevant experience, or a PhD with 2+ years of experience.
• Proficient in modern C, C++, and Python.
• Deep experience in neural network deployment, quantization, and model compression (specifically for Transformers/LLMs).
• Solid understanding of neural network inference frameworks for embedded systems (e.g., QNN, TFLite, NCNN, ONNX).

• 8+ years of experience in embedded Linux development or AI/ML application engineering.
• Experience in video/image processing, computer vision, or multimedia algorithm development (relevant for CLIP/Vision tasks).
• Hands-on experience with LLM/VLM model pipelines, including fine-tuning, evaluation, and optimization on NPU/DSP.
• Experience with Qualcomm AI Stack specifically "AI Engine Direct SDK (QNN)" and "AI Model Efficiency Toolkit (AIMET)".
• Familiarity with hardware accelerators (Hexagon DSP) and embedded architectures.