多模态算法工程师-搜索
Core
Building world-class multimodal search systems for Douyin and international short video products using frontier machine learning and large model technologies.
Role type
Senior IC multimodal algorithm engineer (search)
Builds
Multimodal search systems, AI Q&A/Agents, and large-scale visual search features for short video ecosystems.
Domain
Internet / Search / Multimodal AI
Deliverable
production ML models
Required skills
VLM, LLM, CV, OmniModel, Agent, DeepResearch, visual search algorithms, large-scale contrastive learning, general Tokenizer, SFT, RL, causal inference
Preferred skills
Experience with scaling laws, agentic technology upgrades, automated labeling, generative retrieval, dense personalization, RL-based GE reranking
Technologies
VLM, LLM, Continue Train, SFT, RL, OmniModel, DeepResearch
Responsibilities
Research and apply VLMs and large-scale contrastive learning for high-quality image-text captioning and multimodal representation; Iterate Planner and Summary capabilities for AI Q&A/Agents using SFT and RL; Upgrade relevance, recall, and ranking algorithms for large-scale visual search with generative and dense personalization techniques.
Seniority
Senior, hands-on IC
