Senior AI Engineer, AI Lab

Overview

In this role, you will help shape how AI transforms Economist journalism by fine-tuning LLMs to match The Economist’s editorial voice, building data pipelines for model training, and developing multi-modal generation workflows. You’ll work closely with editorial, product, and design teams to ship AI-powered tools and demos that influence how millions experience our journalism. You will prototype retrieval and audio-first features, while evaluating quality with editorially-grounded metrics. This is a hands-on, high-impact position in a small, ambitious AI Lab.

Pay / Benefits

• 竞争力强的养老金/401(k)计划
• 私立健康保险
• 24/7 心理健康与福利资源
• Work From Anywhere（远程工作）至多25天/年
• 充足的年假与育儿假
• 志愿服务日与搬家假期安排自由度

Responsibilities

• Fine-tune LLMs for tone and style alignment with editorial voice
• Design, curate, and manage fine-tuning datasets with versioning and annotation workflows
• Build and evaluate RAG pipelines with retrieval from structured content
• Prototype and test TTS pipelines for audio-first products (e.g., ElevenLabs, OpenAI TTS)
• Collaborate with infra, frontend, and design to ship internal tools and demos showcasing GenAI capabilities
• Own generation quality evaluation pipelines (BLEU, ROUGE, editorial scoring, custom evals) and human-in-the-loop feedback
• Support product experiments involving real-time generation, summarisation, or personalisation
• Partner with journalists/editors to develop evaluation metrics capturing The Economist’s tone and style

Key requirements

• 3+ years experience building with LLMs or NLP pipelines (OpenAI, Claude, Cohere, Gemini, Mistral, HuggingFace)
• Experience with supervised fine-tuning (SFT), prompt tuning, or instruction tuning on proprietary datasets
• Strong understanding of fine-tuning paradigms, including RLHF/RLAIF
• Strong Python skills, including LangChain, HuggingFace Transformers, and data pipelines (Pandas, DVC, Weights & Biases)
• Experience defining and tracking generation quality with eval metrics like BLEU, ROUGE, editorial scoring
• Exposure to STT / TTS tools for prototyping (Whisper, ElevenLabs, Bark)
• Strong communication and collaboration mindset with editorial and product stakeholders
• Curious and exploratory mindset, comfortable with ambiguity and pushing GenAI boundaries

• 强沟通与跨团队协作
• 好奇心驱动的探索性思维
• 在不确定情形下的自我驱动能力
• LLM/NLP fine-tuning and prompt engineering
• RAG and retrieval pipelines
• Python 编程及数据管线

…