Если вы раньше входили через Google, сбросьте пароль для своей Gmail-почты через кнопку «Забыли пароль?» на экране входа. Затем войдите по email и новому паролю.
Если аккаунта ещё нет, зарегистрируйтесь с Gmail-почтой, после подтверждения почты мы предложим задать пароль.
Что нового
Загружаю обновления...
Что нового
Загружаю обновления...
Работа найдется быстрее с подпискойКандидат найдётся быстрее с подпиской
Чтобы адаптировать резюме под вакансию или составить сопроводительное письмо, загрузите резюме
описание
The Economist Group is a global media and information-services company that delivers trusted analysis and insights to individuals and organisations in more than 170 countries. Its AI Lab explores how generative AI can shape the future of Economist journalism.
задачи
Fine-tune large language models for style and tone alignment with The Economist’s editorial voice;
Design, curate, and manage fine-tuning datasets, including versioning and annotation workflows;
Build and evaluate RAG pipelines incorporating retrieval from structured content;
Prototype and test text-to-speech pipelines for audio-first products;
Collaborate with infrastructure, frontend, and design leads to ship internal tools and demos exploring GenAI capabilities;
Own quality evaluation pipelines, including BLEU, ROUGE, editorial scoring, custom evaluations, and human-in-the-loop feedback loops;
Support product experiments involving real-time generation, summarisation, or personalisation;
Partner with journalists and editors to develop evaluation metrics capturing The Economist’s tone and style.
требования
3+ Years of experience building LLM or NLP pipelines;
Experience with supervised fine-tuning, prompt tuning, or instruction tuning on proprietary datasets;
Strong understanding of fine-tuning paradigms, from SFT to the principles behind RLHF/RLAIF preference modelling;
Strong Python skills with LangChain, HuggingFace Transformers, and data pipelines using Pandas, DVC, and Weights & Biases;
Ability to define and track generation quality using BLEU, ROUGE, and editorial-specific evaluators;
Experience with STT/TTS tools for prototyping;
Strong communication and collaboration skills with editorial and product stakeholders;
Comfortable working in ambiguity and pushing the boundaries of GenAI capabilities;
Nice to have: Experience fine-tuning models for style or persona, exposure to LangSmith or similar observability tools, knowledge of voice synthesis or cloning, experience in media or content production, familiarity with multimodal generation, contributions to open-source LLM tools or libraries, interest in the ethical, editorial, and philosophical questions raised by AI-generated content.
условия
Hybrid working pattern with 3+ days of office attendance required;
Work From Anywhere program allows working from any location where legally permitted for up to 25 days per year;
Highly competitive pension or 401(k) plan;
Private health insurance;
24/7 Counselling and wellbeing resources through the Employee Assistance Program;
Generous annual and parental leave;
Dedicated days off for volunteering and moving home;
Free access to all The Economist content, including online subscription, apps, and podcasts.