Applied AI Researcher – Foundation Models
Pubblicato il 14-09-2026 - Translated in Roma
Overview
As a Deep Learning Scientist in the Foundation Models team, you will drive open multilingual LLM pre-training research from data curation to model evaluation. You’ll design scalable experiments, explore architectures and optimization strategies, and develop multilingual training methods to advance open models beyond English. You’ll work closely with researchers and engineers to translate hypotheses into large-scale experiments on GPU/HPC infrastructure. This role offers the chance to shape Translated’s next generation of multilingual language models and contribute to state-of-the-art benchmarks.
Retribuzione / Benefits
- flexible smart-working policy
- Rome headquarters with on-site facilities
- health-oriented initiatives (massages, sauna, gym, swimming pool)
- psychological, legal, and financial support
- growth-oriented learning paths and networking events
- fully equipped kitchens and cafeteria
Responsabilità
- design and conduct LLM pre-training research
- plan and run scalable experiments, implement and analyze results
- investigate architectures, optimization, training dynamics, and scaling behavior
- develop multilingual training strategies
- conduct data selection, curation,
and mixture design for pre-training
- evaluate models on multilingual and general benchmarks
- monitor state-of-the-art LLMs and benchmark progress
- execute experiments on large-scale GPU and HPC infrastructure
- translate research hypotheses into measurable experiments for large-scale training
Requisiti fondamentali
- 3+ years of relevant experience in Deep Learning, ML, NLP, or LLMs
- strong understanding of modern deep learning and transformer models
- excellent Python programming
- experience designing and running ML experiments
- familiarity with GPU-based training and Unix/Linux systems
- ability to analyze results and make research decisions empirically
- interest in large-scale experimental research and language model pre-training
- ability to read and reproduce recent scientific literature
- excellent written and spoken English
- ability to collaborate effectively with researchers and engineers
- collaboration
- analytical thinking
- clear communication
- transformer-based language models
- Python programming
- GPU-based training