GraemeNail

United Kingdom

Research Engineer in Artificial Intelligence with over 12 years of experience driving research and software development for both on-device inference and large-scale training.

Experience

Research Engineer

Meta, London, UK

Advancing foundation language models (Llama 3 and 4) at Meta Superintelligence Labs. Enhancing pretraining efficiency across distillation, pruning, and scaling, alongside contributions to reasoning and inference optimization. Curating high-quality multilingual web data, including rigorous cleaning and deduplication efforts. [PyTorch, PySpark, Python, C++]

Senior Machine Translation Scientist

Efficient Translation Ltd., UK

Delivered client-specific, high-performance machine translation systems that run locally to preserve privacy. Engineered complete model pipelines, from data bitexting and augmentation through to knowledge distillation, quantization, and evaluation. [C++, Python, HPC]

Postdoctoral Research Associate (StatMT)

The University of Edinburgh, UK

Scaled machine translation pipelines to handle 12 petabytes of web data for the HPLT project. Published open-source bitexts and efficient translation models. Contributed CPU/GPU inference optimizations, achieved competitive results in the WMT Efficiency Tasks. [C++, Python, Bash]

Postdoctoral Research Associate (PPT)

The University of Edinburgh, UK

Enhanced simulations of QCD phenomena for the LHC by improving modeling code for the Herwig and HEJ Monte Carlo frameworks, including the implementation of key computations for vector-boson scattering in the high-energy limit. [C++, Python, Mathematica]

Education

PhD. Theoretical Particle Physics

The University of Manchester, UK

Conducted analytical research and development on computational simulations of QCD phenomenology for the Large Hadron Collider, investigating novel accuracy improvements (KrkNLO) and analyzing parton shower uncertainties.

MSci. in Mathematics and Physics

Durham University, UK

Awarded First Class Honours.

Skills

Technical • C++, Python, Bash, Git, SQL, LaTeX\LaTeX, HTML, CSS, JavaScript, Linux, MacOS

Other • English (Native), Italian (Basic),

Publications

The Llama 3 Herd Of Models (2024)

  • The Llama Team

A New Massive Multilingual Dataset For High-Performance Language Technologies (2024)

  • Ona De Gibert,

  • Graeme Nail,

  • Nikolay Arefyev,

  • Marta Bañón,

  • Jelmer Van Der Linde,

  • Shaoxiong Ji,

  • Jaume Zaragoza-Bernabeu,

  • Mikko Aulamo,

  • Gema Ramírez-Sánchez,

  • Andrey Kutuzov,

  • Sampo Pyysalo,

  • Stephan Oepen,

  • Jörg Tiedemann

OpusCleaner And OpusTrainer, Open Source Toolkits For Training Machine Translation and Large Language Models (2023)

  • Nikolay Bogoychev,

  • Jelmer van der Linde,

  • Graeme Nail,

  • Barry Haddow,

  • Jaume Zaragoza-Bernabeu,

  • Gema Ramírez-Sánchez,

  • Lukas Weymann,

  • Tudor Nicolae Mateiu,

  • Jindřich Helcl,

  • Mikko Aulamo

HPLT: High Performance Language Technologies (2023)

  • Mikko Aulamo,

  • Nikolay Bogoychev,

  • Shaoxiong Ji,

  • Graeme Nail,

  • Gema Ramírez-Sánchez,

  • Jörg Tiedemann,

  • Jelmer Van Der Linde,

  • Jaume Zaragoza

Edinburgh’s submission to the WMT 2022 Efficiency task (2022)

  • Nikolay Bogoychev,

  • Maximiliana Behnke,

  • Jelmer Van Der Linde,

  • Graeme Nail,

  • Kenneth Heafield,

  • Biao Zhang,

  • Sidharth Kashyap

Efficient Machine Translation with Model Pruning and Quantization (2021)

  • Maximiliana Behnke,

  • Nikolay Bogoychev,

  • Alham Fikri Aji,

  • Kenneth Heafield,

  • Graeme Nail,

  • Qianqian Zhu,

  • Svetlana Tchistiakova,

  • Jelmer Van der Linde,

  • Pinzhen Chen,

  • Sidharth Kashyap,

  • Roman Grundkiewicz