Senior LLM Engineer_FedGPT

Taiwan AI Labs is a dynamic startup environment offering exceptional career development opportunities. We invite talented professionals to join us in shaping the future of artificial intelligence (AI) experiences. We are seeking results-driven, highly organized individuals with strong leadership skills to plan, execute, and manage various engineering projects. At Taiwan AI Labs, our mission is clear: to define the future of AI experiences. We are developing LLM-powered agent systems that interact with tools, retrieve information, and perform multi-step reasoning in real-world scenarios. We are looking for a Senior LLM Engineer who can not only build such systems, but also improve model behavior through techniques such as supervised fine-tuning, evaluation, and reinforcement learning to achieve reliable performance in complex tasks. This role involves owning system-level decisions and driving improvements from experimentation to production. 【Responsibilities】 1.LLM Behavior & Reasoning Design and improve LLM-based behavior for complex tasks through strategies such as task decomposition, context management, tool use, and effective integration of external knowledge.2.Model Optimization Experience in model adaptation techniques, such as fine-tuning, preference optimization, or related methods (e.g., reinforcement learning)Make informed trade-offs between prompting, retrieval, data curation, and model adaptation.3.Evaluation & Iteration Design and implement evaluation methods for LLM and agent performance, including task success, reasoning quality, and robustness.Build evaluation datasets and benchmarks aligned with real-world scenarios.Use evaluation results and error analysis to guide improvements across model, data, and system design.4.System Collaboration Work with backend and platform engineers to deploy and iterate on LLM systems.Ensure solutions are practical in terms of latency, cost, and reliability.Contribute to API-based workflows and deployment processes when needed. 【Essential Qualifications】 Experience building and improving LLM applications, such as RAG, agents, or similar systems.Solid foundation in machine learning and deep learning, with understanding of modern LLM techniques.Hands-on experience training or adapting LLMs, including dataset design, fine-tuning, or preference optimization.Ability to design model improvement strategies and make informed trade-offs between prompting, retrieval, data curation, and model adaptation.Experience designing evaluation strategies or frameworks for LLM or NLP systems.Proficiency in Python and modern ML tooling.Ability to translate research ideas into practical improvements that can be validated and deployed in production.Strong ownership in driving ambiguous problems end-to-end, with the ability to collaborate across ML, backend, and product teams and communicate technical trade-offs clearly. 【Preferred Qualifications】 Experience applying reinforcement learning or preference optimization to improve LLM behaviorExperience working with production ML systems, without necessarily owning infrastructure.Working knowledge of Docker, Kubernetes, APIs, or cloud environments in production ML workflows. 【Why Join Us】 Work on real-world AI systems with complex reasoning and decision-making challenges.Improve model behavior through data, evaluation, learning, and applied system design.Drive end-to-end iteration from experimentation to production impact.