Sergey Edunov

Sergey Edunov is an AI research engineering leader whose work spans distributed graph systems, machine translation and language models. He earned an MS in applied physics and mathematics at Moscow Institute of Physics and Technology in 2005, and held software roles at Deutsche Bank, Odnoklassniki and Zynga through 2014. In AI, he worked on Facebook's Apache Giraph infrastructure, coauthored fairseq and multilingual translation research, and coauthored Llama 2. He led Meta's development of Llama 2 and Llama 3 before joining Genesis Molecular AI as senior vice president of foundation models in December 2025; as of October 2026, he is its CTO.

Now at: Genesis Molecular AI
Works: 38
Badges: Bridge

Publication numbers from OpenAlex, as of 22 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

1999–2005

An MS at MIPT

He earns an MS in applied physics and mathematics at the Moscow Institute of Physics and Technology in 2005.
2009–2014

Software engineering roles

From 2009 to 2011, he is an assistant vice president at Deutsche Bank. He works as a software engineer at Odnoklassniki from 2011 to 2012, then as a senior software engineer at Zynga from 2012 to 2014.
2016

Distributed graph infrastructure

He works on distributed graph infrastructure for Apache Giraph, which supports growth and engagement at Facebook. In 2016, he joins the project's project management committee.
2018

Faster neural machine translation

In 2018, he coauthors “Scaling Neural Machine Translation,” which reaches 29.3 BLEU on WMT’14 English-German after 85 minutes of training on 128 GPUs.
2019

Building fairseq

He coauthors fairseq in 2019, an open-source sequence-modeling toolkit for translation, summarization, language modeling and text generation. It supports distributed training across multiple GPUs and machines.
2021–2022

Translation across languages

In 2021, he coauthors CCMatrix, which mines 10.8 billion parallel sentences across 90 languages. In 2022, he coauthors No Language Left Behind, a machine translation project aimed at breaking the 200-language barrier.
2023–2024

Llama and long-context models

In 2023, he coauthors Llama 2, a family of pretrained and fine-tuned language models ranging from 7 billion to 70 billion parameters. In 2024, he coauthors a study that trains long-context LLMs with windows up to 32,000 tokens.
As of October 2026

CTO at Genesis Molecular AI

In December 2025, Genesis appoints him senior vice president of foundation models after he leads development of Meta's Llama 2 and Llama 3 systems. As of October 2026, he is CTO at Genesis Molecular AI, where AI researchers and simulation experts work with drug hunters to optimize drug discovery.