OpenAIResearch Engineer, Frontier Evals & Environments
San Francisco · 全职 · 算法与研究
需要对标的能力
大模型原理 · LLM 基础幻觉治理 · 输出可靠性Agent · 智能体设计工具调用 · Function Calling多 Agent 协作模型微调 · Post-Training评测 · Eval 方法论
岗位职责与要求
About the team
The Agent Post-Training team creates the frontier agents OpenAI ships to the world. We are training the models behind our agents in Codex, ChatGPT, the API, and other frontier products: persistent, proactive intelligence that can operate computers, collaborate with people and other agents, and expand what people and organizations can imagine, attempt, and achieve.
We define what the next generation of agents should be able to do, build the training signal that teaches those abilities, and run the experiments that make them real. Our work spans coding, tool use, computer use, multi-agent coordination, long-horizon execution, factuality, instruction following, calibrated reasoning, and taste.
Our team is where new model capabilities get made. We build the data, environments, graders, training methods, and feedback loops that shape what OpenAI's next agents can do, then carry those capabilities through major training runs and into the products people use.
About the Role
As a researcher working on Frontier Evals & Environments, you will help build north star model environments to drive progress towards safe AGI/ASI. Your work will directly guide the research programs of the most ambitious training runs happening at OpenAI. Some prior open-sourced evaluations built by researchers in this role include GDPval https://openai.com/index/gdpval/, SWE-bench Verified https://openai.com/index/introducing-swe-bench-verified/, MLE-bench https://openai.com/index/mle-bench/, PaperBench https://openai.com/index/paperbench/, and SWE-Lancer https://openai.com/index/swe-lancer/. If you are interested in feeling firsthand the fast progress of our models, and steering them towards good outcomes, this is the role for you.
You will work with researchers, engineers, product teams, infrastructure teams, and safety/alignment partners to decide what should go into major model runs, measure whether it worked, and ship improvements into products used by real people. This is a high-agency role for people who want their work to land directly in frontier models.
In this role, you'll:
- Create ambitious RL environments to push our models to their limits, and measure frontier model capabilities, skills, and behaviors
- Develop new methodologies for automatically exploring the behavior of these models
- Dive deep into the science of measurement, including understanding scalability, reliability, and variance of our evaluation methodology
- Help steer training for our largest training runs, and see the future first
- Design scalable systems and processes to support continuous evaluation
- Build self-improvement loops to automate model understanding
You might thrive in this role if you
- Have strong technical fundamentals in machine learning, software engineering, systems, statistics, or a related field, and can learn quickly across the parts you have not worked in before.
- Have hands-on experience with LLMs, RL, RLHF/RLAIF, post-training, evals, graders, synthetic data, model training, coding agents, tool-using agents, or production ML systems.
- Are excited by open-ended problems where the path is unclear, the signal is noisy, and the right answer requires both research taste and engineering execution.
- Care about product impact and model behavior, not just benchmark movement. You have opinions about what makes an agent useful, reliable, honest, tasteful, and easy to work with.
- Can move from a vague behavioral problem to a concrete experiment: define the hypothesis, build the pipeline, run the model, analyze the result, and decide what to do next.
- Are comfortable working across research, product, infrastructure, data, evals, and safety boundaries, and can communicate clearly with each group.
- Like building load-bearing systems and processes when that is what the team needs, even if the work is not glamorous.
- Want to train and ship the models that make agents genuinely useful for developers, enterprises, researchers, and everyday users.
To notify OpenAI that you believe this job posting is non-compliant, please submit a report through this form https://form.asana.com/?d=57018692298241&k=5MqR40fZd7jlxVUh5J-UeA. No response will be provided to inquiries unrelated to job posting compliance.
At OpenAI, we believe artificial intelligence has the potential to help people solve immense global challenges, and we want the upside of AI to be widely shared. Join us in shaping the future of technology.
查看中文参考译文
关于团队
Agent Post-Training 团队打造 OpenAI 向世界推出的前沿 agent。我们正在为 Codex、ChatGPT、API 及其他前沿产品中的 agent 训练模型：这是一种持久、主动的 intelligence，能够操作计算机、与人及其他 agent 协作，并拓展个人与组织能够想象、尝试和实现的目标。
我们定义下一代 agent 应该具备的能力，构建教授这些能力的 training signal，并运行实验将这些能力变为现实。我们的工作涵盖 coding、tool use、computer use、multi-agent coordination、long-horizon execution、factuality、instruction following、calibrated reasoning 和 taste。
我们的团队是打造新模型能力的地方。我们构建 data、environments、graders、training methods 和 feedback loops，塑造 OpenAI 下一代 agent 的能力，然后将这些能力通过 major training runs 带入人们使用的产品中。
关于职位
作为从事 Frontier Evals & Environments 的研究员，你将帮助构建 north star model environments，以推动向安全 AGI/ASI 的进展。你的工作将直接指导 OpenAI 最雄心勃勃的 training runs 的研究项目。该角色研究员此前构建的一些开源评估包括 GDPval https://openai.com/index/gdpval/、SWE-bench Verified https://openai.com/index/introducing-swe-bench-verified/、MLE-bench https://openai.com/index/mle-bench/、PaperBench https://openai.com/index/paperbench/ 和 SWE-Lancer https://openai.com/index/swe-lancer/。如果你希望 firsthand 感受我们模型的快速进步，并引导它们走向良好结果，这就是适合你的角色。
你将与研究员、工程师、产品团队、基础设施团队以及 safety/alignment 合作伙伴合作，决定什么应该进入 major model runs，衡量是否有效，并将改进交付给真实用户使用的产品。这是一个高自主性的角色，适合希望自己的工作直接落地于 frontier models 的人。
在这个职位上，你将：
- 创建有雄心的 RL environments，将我们的模型推向极限，并衡量 frontier model capabilities、skills 和 behaviors
- 开发自动探索这些模型行为的新方法
- 深入研究 measurement 科学，包括理解我们评估方法的可扩展性、可靠性和方差
- 帮助引导我们最大规模 training runs 的 training，并 first 看到未来
- 设计可扩展的系统和流程以支持 continuous evaluation
- 构建 self-improvement loops 以实现 model understanding 的自动化
如果你具备以下条件，你可能在这个角色中茁壮成长
- 在 machine learning、software engineering、systems、statistics 或相关领域有扎实的技术基础，并能快速学习之前未涉及的领域。
- 具有 LLMs、RL、RLHF/RLAIF、post-training、evals、graders、synthetic data、model training、coding agents、tool-using agents 或 production ML systems 的实践经验。
- 对路径不明、信号嘈杂、正确答案需要同时兼具 research taste 和 engineering execution 的开放式问题感到兴奋。
- 关注 product impact 和 model behavior，而不仅是 benchmark movement。你对什么使 agent 有用、可靠、诚实、有品位且易于合作有自己的见解。
- 能够从模糊的行为问题推进到具体实验：定义假设、构建 pipeline、运行模型、分析结果，并决定下一步行动。
- 能够跨越 research、product、infrastructure、data、evals 和 safety 的边界工作，并能与每个团队清晰沟通。
- 在团队需要时乐于构建承重的系统和流程，即使这项工作并不光鲜。
- 希望训练并交付使 agent 真正对 developers、enterprises、researchers 和 everyday users 有用的模型。
如需通知 OpenAI 你认为此职位发布不合规，请通过此表单提交报告 https://form.asana.com/?d=57018692298241&k=5MqR40fZd7jlxVUh5J-UeA。与职位发布合规性无关的询问将不予回复。
在 OpenAI，我们相信 artificial intelligence 有潜力帮助人们解决巨大的全球挑战，我们希望 AI 的益处能够被广泛分享。加入我们，共同塑造技术的未来。
译文仅供辅助阅读，具体要求以原始招聘来源为准。