AI Researcher – Multilingual Data

About the Role

We are seeking an AI Researcher specializing in multilingual data to join a pioneering research environment in Germany. This role involves spearheading the creation and optimization of high-quality multilingual datasets and developing innovative research methodologies that enhance language models across various languages and fields. This position bridges research and engineering, with a goal to translate groundbreaking findings into scalable production systems. The environment is fast-paced and innovation-focused, offering opportunities to publish research and collaborate with leading experts to advance multilingual AI globally.

Key Responsibilities

• Conduct research concentrating on multilingual datasets, encompassing tasks such as data collection, filtering, deduplication, quality evaluation, and dataset improvement.
• Implement novel strategies for addressing low-resource and infrequently represented languages through sophisticated sampling methods, data augmentation, and curriculum learning.
• Enhance multilingual large language models by improving cross-lingual transferability, alignment, robustness, and representational learning.
• Create and manage multilingual evaluation benchmarks to accurately gauge model performance across different languages.
• Collaborate closely with machine learning engineers and research professionals to influence model training workflows, architectures, and production implementations.
• Disseminate research insights at top-tier AI and NLP conferences and contribute to open-source projects where applicable.
• Apply research results practically to improve AI systems ready for production deployment.

Required Qualifications

• Strong academic and practical background in Natural Language Processing, Machine Learning, Artificial Intelligence, or closely related disciplines.
• Demonstrable experience in multilingual or cross-lingual language modeling with published work in reputable conferences or journals such as ACL, EMNLP, NeurIPS, ICML, or ICLR.
• Proficiency in handling large-scale multilingual text datasets and familiarity with contemporary machine learning pipelines.
• Deep understanding of multilingual tokenization, vocabulary strategies, transfer learning, representation learning techniques, dataset evaluation, filtering, and mitigation of bias.
• Expertise in Python programming and deep learning frameworks like PyTorch or JAX.
• Self-driven with the capability to manage research projects and deliver results efficiently within a dynamic startup environment.
• Experience with low-resource languages, non-Latin scripts, multilingual evaluation frameworks such as XTREME, FLORES, TyDi QA, or involvement in open-source NLP endeavors and large language model training is highly desirable.

Compensation and Benefits

• Attractive remuneration package commensurate with experience.
• Equity participation in an early-stage, fast-growing organization.
• Significant influence on research directions and technical choices.
• Balanced opportunities to engage in academic research alongside impactful real-world applications.
• Access to extensive multilingual datasets, state-of-the-art AI infrastructure, and rapid iterative experimentation.
• A collaborative culture promoting innovation, excellence in research, and ongoing professional growth.
• Chance to present work at leading international AI and NLP conferences and contribute to influential open-source projects.

Additional Information

This role is operated by a partner company located in Germany, which will handle application reviews and subsequent recruitment steps independently. Candidates will engage directly with the hiring organization after shortlisting. The application and selection system employs AI technology for efficient and unbiased candidate screening; however, hiring decisions remain with human recruiters.

Applicants should note the data privacy details related to the hiring process under GDPR compliance, with full rights to access, correct, or delete personal data.