Responsibilities
Team introduction: ByteDance's search team is mainly responsible for the innovation and architecture research and development of search algorithms for products such as Douyin, Toutiao, Xigua Video, etc. We use cutting-edge machine learning technology for end-to-end modeling and constantly innovate and make breakthroughs. At the same time, we focus on the construction and performance optimization of distributed systems and machine learning systems, from memory, Disk and other optimization to the exploration of algorithms such as index compression, recall, and sorting, fully providing students with opportunities to grow themselves. The main working directions include: 1) Exploring the most cutting-edge NLP technology: from basic word segmentation, NER, to application query analysis, basic correlation, etc., the full link application deep learning model is challenging 2) Exploring cross-modal matching technology: applying CV+NLP deep learning technology in search, so that video search has stronger retrieval capabilities 3) Exploring large-scale streaming machine learning technology: applying large-scale machine learning to solve recommendation problems in search, making search more personalized and understand you 4) Exploring the architecture of a data scale of hundreds of billions of yuan: from large-scale offline computing, the performance and scheduling optimization of distributed systems, to building high availability, high throughput and low latency online services, all aspects are carried out. Involve research and innovation. 1. Participate in search engine research and development, explore personalized behavior modeling of the entire search link (analysis, recall, rough arrangement, fine arrangement, mixed arrangement), including CTR, CVR estimate, vector recall, value arrangement, RAG, NLP, LLM, multi-modal, machine learning, deep learning, etc., to promote the implementation and improvement of search algorithms in international e-commerce scenarios, and improve the search experience of 100 million users 2. Participate in the optimization and iteration of international e-commerce search algorithms, improve the conversion efficiency, user experience and supply ecosystem solve technical problems such as multilingual correlation matching, authoritative perception, grass planting content understanding, repeated distribution, and counterfeit goods governance, and ultimately optimize the basic search quality of various e-commerce business forms such as content e-commerce and traditional shelf e-commerce to ultimately improve the shopping conversion efficiency of goods, grass planting videos and live streaming, and promote GMV growth 3. Deeply participate in the demand design of core search products, be responsible for the high-quality delivery of algorithms and projects, and continuously optimize the results and improve product experience 4. Mining data, building Query understanding, recall, sorting and other models, and improving e-commerce search algorithm capabilities 5. Learn cutting-edge technologies and explore the implementation of innovative technologies such as big models in AI search scenarios.
Qualifications
1. Bachelor degree or above, computer, electronics, mathematics and other related majors 2. In-depth researchers are preferred in one or more fields such as search, recommendation, search and promotion, computer vision CV, natural language processing NLP, natural language understanding NLU, multimodal, machine learning, deep learning, etc. 3. Familiar with the Linux development environment, proficient in C++ and Python language, familiar with Html and related technologies, and familiar with network programmers are preferred 4. Have good problem analysis and solving skills, communication and collaboration skills, proactive work, able to cooperate harmoniously with the team, explore new technologies and promote technological progress. Bonus: 1. Have excellent basic algorithms and solid machine learning/deep learning foundations, be familiar with technologies in CV, NLP, RL, ML and other fields, and publish papers in top conferences/journals such as ICML, ACL, COML, EMNLP, CVPR, ECCV, ICCV, NeurIPS, ICLR, SIGGRAPH or SIGGRAPH Asia 2. Have excellent code skills, proficient in C/C++ or Python programming languages, and winners of ACM/ICPC, NOI/IOl, Top Coder, Kaggle and other competitions are preferred 3. In LLM, multimodal, large model, basic model, world model, RL and other fields, those who dominate the overly influential projects are preferred 4. Those with experience in large-scale search engines, recommendation systems, distributed systems, computing advertising, ultra-large-scale data computing, etc. are preferred.
