About this role
Build core Agent logic (task planning, orchestration, tool calling, multi-turn dialogue, memory, RAG, context engineering, and multi-agent collaboration), lead continuous pre- and post-training including dataset and RL reward model design, develop memory architectures and multimodal RAG systems, integrate tool layers and external APIs, design prompts and evaluation systems, and monitor product metrics to optimize Agents and models.
Skills for this role
Task planningOrchestrationTool callingMulti-turn dialogue managementMemory architecturesRAGContext engineeringMulti-agent collaborationContinuous pre-trainingPost-trainingDataset creationData pipelinesRL reward model designInstruction followingReasoningPersonalized dialogueMultimodal conversationLong-term memoryShort-term memoryRecallRankingLong-text processingMulti-document synthesisAPI integrationMCP integrationSearch integrationCode interpreter integrationBrowser integrationSandbox integrationPrompt design and tuning