面議(經常性薪資達4萬元或以上) 台北市大同區 3年工作經驗 1天前更新
About the Role:
You will join us as an AI engineer who works across several AI projects in the group. You will be part of the team led by our Tech Lead. One of them is our new humanoid robot lab in Taipei, where we explore how robots can take over repetitive manual work. Other projects bring AI agents, large language models (LLMs) and knowledge systems into software used by our group companies. You will build things that work, and you will help your colleagues learn to build them too. Growing our team’s AI skills is part of the job, not an extra.
What You Will Do
• Design, build and ship AI features in real software: LLM agents, tool use, retrieval-augmented generation (RAG), speech and vision
• In the robot lab, help the robot understand and act — speech interaction, object detection, simple pick-and-place — and run models on edge devices such as NVIDIA Jetson
• Choose the right approach — ready-made models, APIs or fine-tuning — and explain the trade-offs in plain words
• Measure quality, speed and cost, and report honestly what works, what does not, and why
• Turn what you learn into reusable components and clear documentation that other projects can use
• Run workshops and coach colleagues who are new to AI
• Work in English with colleagues in Germany
Requirements:
• Typically 3+ years in software or machine learning engineering, with AI features in real use
• Strong Python, and experience with at least one machine learning framework such as PyTorch
• Hands-on work with LLM applications: agents, tool use, RAG, prompt design and evaluation
• Depth in at least one area: computer vision, speech, or robotics
• Experience deploying models on Linux — containers, APIs, cloud or edge
• You can explain technical ideas to people who are not engineers, and you enjoy doing it.
• You are comfortable with uncertainty — some of this work is new for us, too
• You can work on-site in our Taipei lab several days a week; robot hardware cannot be handled remotely
Nice to Have
• ROS 2 (Robot Operating System 2)
• Robot manipulation, imitation learning or reinforcement learning
• Vision-Language-Action (VLA) models or robot foundation models
• Simulation tools such as NVIDIA Isaac Sim, Omniverse or MuJoCo
• Model fine-tuning and MLOps (machine learning operations)
• C++
• Experience in logistics, warehousing or retail
展開