{"id":3509399,"url":"https://alion.io/job/unity-omoxingpingcegongchengshicoding-agenta232277","title":"大模型评测工程师(Coding Agent)(A232277)","company":{"id":2511966,"name":"Unity China","domain":"unity.cn","url":"https://alion.io/company/unity-cn","size_band":null,"is_staffing_agency":false,"employer_type":"direct","is_intermediary":false,"listed_via":null,"ats_vendor":null,"truth_index":null},"role":"AI/ML","role_family":"AI/ML","seniority":null,"employment_type":"full_time","work_mode":"on_site","remote_scope":null,"remote_scope_basis":null,"remote_working_hours":null,"hiring_geo_confidence":"structured","locations":["Beijing, China"],"countries":["CN"],"hiring_countries":[],"hiring_countries_total":0,"salary":{"min":40000,"max":60000,"currency":"CNY","period":"month","gross":null,"usd_annual":108336},"salary_estimate":null,"experience_years_min":null,"visa_sponsorship":false,"relocation_package":false,"has_equity":false,"technologies":[{"name":"DeepSpeed","optional":false},{"name":"DPO","optional":false},{"name":"FSDP","optional":false},{"name":"Function Calling","optional":false},{"name":"GRPO","optional":false},{"name":"LLM","optional":false},{"name":"Megatron-LM","optional":false},{"name":"Post-training","optional":false},{"name":"Python","optional":false},{"name":"Reinforcement Learning","optional":false},{"name":"RLHF","optional":false},{"name":"SFT","optional":false}],"status":"live","first_seen_at":"2026-10-21T06:55:37Z","employer_posted_date":null,"last_verified_at":"2026-10-21T06:55:37Z","board_verified":false,"closed_at":null,"days_open":18,"trust":{"level":"not_scored","repost_count":null,"flags":[],"days_open":18},"description":"我们正在构建面向游戏开发的下一代 AI Coding Agent，探索大语言模型在代码生成、工具调用、环境交互和复杂软件开发任务中的智能化能力。团队聚焦 LLM Post-training、Agentic Reinforcement Learning、Code Agent 和 Agent Evaluation 等前沿方向，希望打造能够自主完成 Unity 游戏开发任务的智能开发助手。\n\n你将重点负责 Unity Coding Agent 的评测体系建设，围绕游戏开发、工具调用与长链路任务设计 Benchmark、任务集、评测指标和自动化验证流程；研究并落地 LLM/Agent Evaluation 前沿方法；基于真实 Unity 工程和大规模 Agent Trajectory 推动“数据生成—评测—训练”闭环，用评测结果反向驱动数据构造、模型后训练和 Agent 能力持续提升。\n\n岗位职责\n\n- 负责构建 Unity Coding Agent 的自动化评测体系，围绕游戏开发、工具调用与长链路任务，设计 Benchmark、任务集、评测指标和自动化验证流程，覆盖 Agent 规划、执行、工具使用和最终产物质量。\n\n- 研究并落地 LLM/Agent Evaluation 前沿方法，包括基于规则、程序执行、视觉理解和 LLM-as-a-Judge 的多维评测方法，分析 Agent 规划、执行、工具调用和失败模式，形成可复用的评测范式。\n\n- 基于真实 Unity 工程和大规模 Agent Trajectory，探索“数据生成—评测—训练”闭环，通过失败案例归因定位数据缺陷和能力瓶颈，用评测结果反向驱动数据构造、模型后训练和 Agent 能力持续提升。\n\n- 建设面向 Coding Agent 的评测数据与任务生产流程，包括任务集构建、测试用例/验证程序生成、Trajectory 采集与质量评估、评测数据配比和迭代优化。\n\n- 与后训练团队协作，将评测洞察转化为 SFT、Preference/RL 数据和训练策略建议，参与模型能力对比、checkpoint 评估、能力回退与遗忘问题分析，推动 Base Model → SFT Model → RL Model → Production Model 的迭代验证。\n\n- 与 Agent 工程团队协作，优化主 Agent + SubAgent 架构中的模型能力分工和协作策略，推动评测平台、自动化验证流程和指标体系的工程化落地。\n\n职位要求\n\n- 本科及以上学历，计算机、人工智能、数学等相关专业，硕博优先。\n\n- 具备大模型评测 / Agent Evaluation 相关经验，深入参与过 Benchmark、任务集、评测指标或自动化验证流程建设。\n\n- 熟悉 LLM/Agent 评测方法，包括基于规则、程序执行、视觉理解、LLM-as-a-Judge 等至少一种多维评测方式，能够设计面向 Agent 规划、执行、工具使用和失败模式分析的评测方案。\n\n- 具备 Agent、Function Calling、Tool Use 或 Code LLM 相关经验，理解多轮 Agent Trajectory 数据生产和自动化评测流程。\n\n- 熟悉 LLM 后训练完整流程，包括 SFT、RLHF、DPO、GRPO 等训练方法，理解评测如何驱动后训练迭代；有实际落地经验优先。\n\n- 具备较强的训练数据工程意识，包括 SFT 数据构造、清洗、过滤和质量控制，Preference/RL 数据设计，多轮 Agent Trajectory 数据生产，以及数据混合比例设计和效果分析。\n\n- 具备较强实验分析能力，能够从数据、训练策略和模型能力多个角度定位问题，并形成可执行的优化建议。\n\n- 具备扎实的工程能力，熟练使用 Python，能够构建自动化评测、数据处理和实验分析工具；了解 DeepSpeed、FSDP、Megatron 等分布式训练框架者优先。\n\n加分项\n\n- 有 Code LLM、Coding Agent 的训练或评测经验。\n\n- 有大规模 Agent Trajectory 数据生产和自动化评测经验。\n\n- 有模型能力分析、Benchmark 构建或数据驱动模型优化经验。\n\n- 熟悉 Unity 游戏开发场景，包括 Shader、粒子、UI、场景构建、性能分析等。\n\n- 有大模型后训练团队实习/工作经历，了解 checkpoint 分析、SFT/RL 模型能力对比、多能力模型融合和效果验证。","description_format":"text","description_chars":1891,"description_truncated":false,"requirements":{"experience_years_min":null,"management_years_min":null,"team_size_min":null,"manages_managers":false,"education":null,"security_clearance":false,"languages":[]},"benefits":[],"hiring_locations":[],"hiring_excludes":[],"relocation_offered":false,"industries":[],"lifecycle":[{"event":"open","at":"2026-10-30T06:37:21Z"}],"visa":[],"liveness":{"score":50,"band":"ok","label":"Likely open","p_open":0.85,"p_active":0.781,"p_room":0.75,"age_days":17,"expected_fill_days":25,"reasons":["seen:17","win:late"],"computed_at":"2026-11-08T05:47:12Z"},"pay":{"stated_usd_annual":108336,"is_top_pay":false},"html_url":"https://alion.io/job/unity-omoxingpingcegongchengshicoding-agenta232277","json_url":"https://alion.io/job/unity-omoxingpingcegongchengshicoding-agenta232277.json","meta":{"generated_at":"2026-11-08T08:37:11Z","cache_seconds":300,"methodology":"https://alion.io/methodology","terms":"https://alion.io/terms","contact":"https://alion.io/contact","api":"https://alion.io/developers","about":"Alion is a live layer of people, companies and AI agents: who they are, whether they are real and active right now, what they do and how to work with them, readable by people and by agents and paid per call.","catalog":"https://alion.io/catalog.json","usage":{"tier":"crawler_verified","counted_by":"address","units_charged":1,"used_today":18671,"day_limit":null,"remaining_today":null,"minute_limit":300,"resets_at":"2026-11-09T00:00:00Z"}},"offers":[{"id":"company.slices","title":"One company in depth, by slice","status":"live","price":{"credits":0.02,"usd":0.002,"plus_per_slice":{"credits":0.05,"usd":0.005}},"unit":"per company, plus each slice with data","note":"the employer in depth","call":{"mcp_tool":"get_company","arguments":{"id":2511966},"rest":"https://alion.io/mcp/rest/get_company?id=2511966"},"human":"https://alion.io/catalog?offer=company.slices&for=job%2Funity-omoxingpingcegongchengshicoding-agenta232277"},{"id":"market.stats","title":"A market slice: pay, demand and time to fill","status":"live","price":{"credits":1,"usd":0.1},"unit":"per slice","note":"pay, demand and time to fill for this role and place","call":{"mcp_tool":"market_stats"},"human":"https://alion.io/catalog?offer=market.stats&for=job%2Funity-omoxingpingcegongchengshicoding-agenta232277"},{"id":"job.search","title":"Open jobs by role, technology, place, pay and visa","status":"live","price":{"credits":0.02,"usd":0.002},"unit":"per posting in a list","note":"similar open postings","call":{"mcp_tool":"search_jobs"},"human":"https://alion.io/catalog?offer=job.search&for=job%2Funity-omoxingpingcegongchengshicoding-agenta232277"},{"id":"company.verify","title":"Is this company real and active right now","status":"pilot","price":null,"unit":"per company","request":{"url":"https://alion.io/catalog/request","method":"POST","body":"{\"offer\": \"company.verify\", \"for\": \"job/unity-omoxingpingcegongchengshicoding-agenta232277\", \"note\": \"what you need it for\"}"},"human":"https://alion.io/catalog?offer=company.verify&for=job%2Funity-omoxingpingcegongchengshicoding-agenta232277"}],"agent":{"mcp":"https://alion.io/mcp/jobs","mcp_open":"https://alion.io/mcp/open","next":[{"tool":"get_job","arguments":{"id":3509399}},{"tool":"search_jobs","arguments":{"company":"unity.cn","detail":"compact"}},{"tool":"match_people","arguments":{"job_id":3509399},"note":"people open to offers who fit this vacancy"}],"note":"The same data through MCP tools: https://alion.io/mcp/open needs no sign-in (compact searches, postings, people); https://alion.io/mcp with an Alion account has everything. Docs: https://alion.io/developers"}}