Research and refine LLM post-training methods and training recipes, designing experiments around target tasks to improve performance and generalization.
Design, build, and optimize end-to-end personalized recommendation systems across multiple user experiences, including personalized content feeds, trending content, and category-based recommendations.
Participate in the development of core capabilities for the advertising system, including AI infrastructure, machine learning and recommendation algorithms, to enhance ad recommendation effectiveness. Analyze foundational data, uncover user interests, and improve the prediction capabilities of the advertising system.
Build core Agent logic, including but not limited to task planning and orchestration, tool calling, multi-turn dialogue management, memory, RAG, context engineering, and multi-agent collaboration.
Participate in the development of core capabilities for the advertising system, including AI infrastructure, machine learning and recommendation algorithms, to enhance ad recommendation effectiveness. Analyze foundational data, uncover user interests, and improve the prediction capabilities of the advertising system.
Responsible for core algorithm models in marketing scenarios, including but not limited to user behavior prediction, user profiling, ROI estimation, and traffic allocation strategies, to improve marketing efficiency and conversion rates.
Design and implement advanced reasoning algorithms for large language models, enabling them to solve complex, multi-step problems, perform logical deductions, and engage in abstract thinking.
Design of Reward Models: In the RL process, designing an effective reward model is crucial. It must accurately reflect the effectiveness of the reasoning process and guide the model to iteratively improve its reasoning ability. This involves not only setting appropriate evaluation criteria across different tasks, but also ensuring the reward model to adapt dynamically during training to match the evolving model performance.
Develop and optimise model quantization algorithms, including post-training quantization (PTQ), quantization-aware training (QAT), and related compression techniques.
Technical Strategy: Define the AIGC roadmap across pre-training and post-training workstreams — from distributed training infrastructure to alignment, evaluation, and inference deployment.
Participate in research of advanced image/video processing algorithms using for autonomous driving, machine vision, and professional photography products.
Conduct research and development on large multimodal models (LMMs) and AIGC generation algorithms, with a focus on image and video generation and editing, continuously advancing key algorithmic capabilities including but not limited to multimodal controllable generation, video continuation, image and video editing, and unified generation-and-understanding models;