Collect first agent research and market sources

2026-07-08 12:08:24 +08:00
parent 36c2499d3c
commit 1cbc98e432
34 changed files with 3417 additions and 58 deletions
@@ -0,0 +1,92 @@
+# Job: 百度 - 2027AIDU Agent 应用全栈工程师
+
+---
+type: job
+company: 百度
+role: 2027AIDU-Agent应用全栈工程师
+location: 北京市
+source_url: https://talent.baidu.com/jobs/list?projectType=3&recruitType=GRADUATE
+source_name: 百度校园招聘
+source_quality: official
+collected_at: 2026-07-08
+posted_at: 2026-05-12
+first_seen_at: 2026-07-08
+last_checked_at: 2026-07-08
+snapshot_type: new
+job_code: J99974
+previous_snapshot:
+status: active
+level: graduate
+team: AIDU项目
+business_area: search / healthcare / enterprise service / data analysis / office automation
+employment_type: campus
+salary:
+skills:
+  - agent
+  - planning
+  - tool-use
+  - function-calling
+  - memory
+  - reasoning
+  - state-management
+  - multi-agent
+  - rag
+  - agent-evaluation
+topics:
+  - workflow-agent
+  - agent-architecture
+  - evaluation
+models:
+  -
+related_papers:
+  - 2026-dialogue-swebench
+  - 2026-swe-evo
+related_experiments:
+  -
+related_projects:
+  -
+relevance: high
+---
+
+## Source Snapshot
+
+- source: 百度官方招聘页
+- collected_at: 2026-07-08
+- availability: 页面可访问，岗位列表显示发布日期 2026-05-12。
+
+## Change Snapshot
+
+- first_seen_at: 2026-07-08
+- last_checked_at: 2026-07-08
+- snapshot_type: new
+- previous_snapshot:
+- changed_fields:
+
+## Raw JD Summary
+
+岗位强调构建下一代 Autonomous Agent 系统，要求从对话工具转向可执行任务的智能体，覆盖 planning、acting、reflection、Tool/API 调用、长短期记忆、多轮推理、状态管理、多 Agent 协作、RAG 融合和 Eval。
+
+## Responsibilities
+
+- 构建具备任务执行能力的 Agent 应用系统。
+- 搭建 Tool/API、memory、reasoning、state 等核心能力。
+- 推动 Agent 在搜索、医疗、企业服务、数据分析和自动化办公等场景落地。
+
+## Extracted Signals
+
+- hard requirements: full-stack Agent engineering, tool/function calling, memory, state, eval。
+- preferred skills: 前沿方向跟踪，包括 Tool Learning、Self-Improvement、World Model、多模态 Agent。
+- hidden signals: 全栈岗位已经不只是 Web 工程，开始要求 Agent runtime 和评估能力。
+- business direction: 企业服务、数据分析、自动化办公。
+
+## Knowledge Links
+
+- concepts: workflow-agent, human-in-the-loop, eval
+- papers: Dialogue-SWEBench, SWE-EVO
+- experiments: workflow-agent prototype
+- projects:
+
+## Gaps for Us
+
+- 需要补一个 Agent runtime / state management 的工程专题。
+- 需要整理 Tool Learning 和 Self-Improvement 的论文清单。