微软Senior Applied Scientist(LLM)
任职要求
• M.S. or Ph.D. in Computer Science, Machine Learning, or a related field, or equivalent practical experience.• 4+ years of experience in applied machine learning, with a focus on LLMs, agent systems, or reinforcement learning.• Strong hands-on experience with model training pipelines using PyTorch, TensorFlow, JAX, or similar frameworks.• Familiarity with distributed training, prompt engineering, evaluation strategies, and model deployment best practices.• Experience with retrieval-augmented generation (RAG), tool use, planning agents, or long-context modeling is a plus.• Solid publication record (e.g., NeurIPS, ICLR, ACL, ICML, EMNLP) is a plus, but emphasis is placed on practical contributions.• Strong coding and debugging skills, and comfort working in cross-functional, agile environments. M…
工作职责
• Design and implement advanced LLM-based architectures and agentic systems for real-world product scenarios.• Lead model training and evaluation efforts, including data preprocessing, fine-tuning, and inference optimization.• Collaborate across teams to deliver robust, scalable models aligned with product objectives and user value.• Apply and adapt research ideas to solve practical challenges in reasoning, planning, memory, and alignment.• Monitor and improve model performance post-deployment through data-driven iteration and error analysis.• Contribute to technical discussions, model reviews, and best practices within the applied science community.
• Build benchmarks, evaluation datasets, metrics, and methods to assess and improve the performance and effectiveness of language models and prompts and drive iterative enhancements. • Deliver high-impact analyses that generate actionable insights to steer product and business decisions and improve user satisfaction. • Collaborate closely with engineering, research, product, and other teams to ensure AI-driven experiences meet quality and user experience standards. • Analyze latest AI innovations and explore opportunities to apply cutting-edge techniques for building scalable, high-impact solutions that enhance product capabilities and deliver exceptional user experiences.
Design, training/evaluation, development/deployment of advanced models (LLM/MLLM/SLM) for ads content creation, ads performance optimization, and automated customer support. Apply strong problem-solving and data analysis skills to identify opportunities, optimize performance, and deliver measurable business impact.Stay up to date with the latest advancements in generative AI, LLM-driven agent, online advertisement, and make integration and improvement for our systems.Develop scalable and robust algorithms and pipelines for efficient model training, data processing, online serving and measurement.Collaborate with cross-functional teams, such as product managers and Infra/UI engineers, to define strategies and deliver solutions that enhance our systems.Present problems, insights, technical solutions, and results clearly to both technical and non-technical stakeholders.
• Develop next-generation AI experiences for Microsoft Edge — leverage machine learning and generative AI to reinvent how users browse, search, and interact with the web. • Advance Edge’s contextual intelligence by building models that synthesize browsing history, page content, and user intent to deliver proactive, personalized, and trustworthy assistance. • Drive innovation in agentic systems, prototyping and productionizing conversational, reasoning, and planning models that transform Edge from a static browser into a true AI companion. • Collaborate cross-functionally with product managers, designers, and engineers to translate AI capabilities into elegant, high-utility user experiences. • Own the full AI feature development lifecycle — from data pipeline and evaluation metric design to model and prompt tuning, quality validation, and continuous improvement.
• Develop next-generation AI experiences for Microsoft Edge — leverage machine learning and generative AI to reinvent how users browse, search, and interact with the web. • Advance Edge’s contextual intelligence by building models that synthesize browsing history, page content, and user intent to deliver proactive, personalized, and trustworthy assistance. • Drive innovation in agentic systems, prototyping and productionizing conversational, reasoning, and planning models that transform Edge from a static browser into a true AI companion. • Collaborate cross-functionally with product managers, designers, and engineers to translate AI capabilities into elegant, high-utility user experiences. • Own the full AI feature development lifecycle — from data pipeline and evaluation metric design to model and prompt tuning, quality validation, and continuous improvement.