AI/ML工程实习生
AI/ML Engineering Intern
位置：远程（美国）
工作模式：远程
行业：AI招聘和人才市场（B2B SaaS）
薪酬：每小时35-50美元
关于公司
TalentPluto 是一家由 Y Combinator 支持的初创公司，正在为高增长公司打造招聘工具链。我们运营两个相互关联的产品：一个人才网络，将工程、市场推广和运营人员安置到快速发展的初创公司；以及一个 AI 招聘工具，允许招聘团队直接从他们的助理中运行候选人搜索、分层丰富、基于评分表的评分和外向活动。我们与多个跨 AI、开发者工具、金融科技和医疗保健领域的风险投资支持的初创公司合作。团队规模小，由创始人领导，实习生参与实际问题，而不是练习问题。
机会
这是一个在产品内部的实用 AI 实习岗位，模型输出即产品。我们使用 LLM 来阅读非结构化的专业背景，从中提取结构，根据职位对候选人进行评分，并对招聘人员看到的内容进行排序。这些步骤中的每一个都可能被错误地执行，因此有趣的工作不仅在于构建流程，还在于评估质量。
你将处理大量混乱数据的 LLM 流程：设计提示和提取逻辑，构建能告诉你更改是否有效的评估，以及改进检索和排名，使当前答案不够好的情况得到改善。这是应用性而非研究性的工作，你将能够看到你所构建内容对真实用户的影响。
职责
- 在真实产品数据上构建和改进 LLM 流程，用于提取、评分和排序。
- 设计衡量输出质量的评估，并用它们来判断更改是否有所改进。
- 处理混乱、高容量的非结构化数据并将其转化为可靠结构。
- 改进检索和排序，使最相关的结果优先显示。
- 调查错误或低质量的输出并追溯其原因。
要求
- 具备 Python、TypeScript、JavaScript 或类似语言的扎实编程基础。
- 对应用 AI、机器学习、LLM、排序、检索或数据系统感兴趣。
- 习惯处理混乱数据并评估输出质量。
- 有构建 AI/ML、数据、课程、研究或副业项目的经历。
- 强大的解决问题能力，愿意快速学习。
- 清晰的沟通者，能够在远程环境中独立工作。
查看英文原文
Location: Remote (United States)
Work Model: Remote
Industry: AI recruiting and talent marketplace (B2B SaaS)
Compensation: $35–$50 per hour
About the Company
TalentPluto is a Y Combinator-backed startup building the hiring stack for high-growth companies. We run two connected products: a talent network that places engineering, go-to-market, and operations people into fast-moving startups, and an AI recruiting tool that lets a hiring team run candidate search, tiered enrichment, rubric-based scoring, and outbound campaigns directly from their assistant. We work with dozens of venture-backed startups across AI, developer tools, fintech, and healthcare. The team is small and founder-led, and interns work on live problems rather than practice ones.
The Opportunity
This is an applied AI internship inside a product where the model output is the product. We use LLMs to read unstructured professional histories, extract structure from them, score candidates against a role, and rank what a recruiter sees first. Every one of those steps can be confidently wrong, so the interesting work is as much about evaluating quality as it is about building the pipeline.
You will work on real LLM workflows against messy data at volume: designing prompts and extraction logic, building the evaluations that tell you whether a change actually helped, and improving ranking and retrieval where the current answer is not good enough. This is applied rather than research work, and you will be able to see the effect of what you build on what real users get.
Responsibilities
- Build and improve LLM workflows for extraction, scoring, and ranking against real product data.
- Design evaluations that measure output quality, and use them to judge whether a change is an improvement.
- Work with messy, high-volume unstructured data and turn it into reliable structure.
- Improve retrieval and ranking so the most relevant results surface first.
- Investigate incorrect or low-quality output and trace it back to its cause.
Requirements
- Strong programming fundamentals in Python, TypeScript, JavaScript, or a similar language.
- Interest in applied AI, machine learning, LLMs, ranking, retrieval, or data systems.
- Comfort working with messy data and evaluating output quality.
- Has built AI/ML, data, class, research, or side projects.
- Strong problem-solving ability and a willingness to learn quickly.
- Clear communicator who can work independently in a remote environment.
Originally posted on Himalayas