Ola Engkvist

Ola Engkvist is a computational chemist who builds machine-learning methods for designing small-molecule drugs. After a PhD in computational chemistry at Lund University and a postdoc at Cambridge University, he joined AstraZeneca in 2004. He coauthored a deep reinforcement learning method for molecular de novo design in 2017, the AiZynthFinder retrosynthesis software in 2020 and the open-source REINVENT 4 framework in 2024, and Clarivate named him a Highly Cited Researcher in 2025. He is Executive Director and Head of Molecular AI in Discovery Sciences at AstraZeneca R&D, a Fellow of ELLIS and an adjunct professor at Chalmers University of Technology.

Now at: AstraZeneca, Chalmers University of Technology and Cambridge Crystallographic Data Centre
Works: 393
Citations: 15,345
h-index: 56
Badges: Rising
Bridge
Wayfarer
Loyal

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

By 2004

From Lund to AstraZeneca

He earns a PhD in computational chemistry at Lund University and does a postdoc at Cambridge University. After working for two biotech companies, he joins AstraZeneca in 2004.
2017

Reinforcement learning for molecule design

In 2017, he coauthors "Molecular De Novo Design through Deep Reinforcement Learning" with Marcus Olivecrona, Thomas Blaschke and Hongming Chen. The method tunes a sequence-based generative model to produce structures with specified properties, and for the dopamine receptor type 2 target more than 95% of the generated structures are predicted to be active.
2020

AiZynthFinder retrosynthesis software

In 2020, he coauthors AiZynthFinder with Samuel Genheden, Amol Thakkar, Veronika Chadimová, Jean-Louis Reymond and Esben Bjerrum. The open-source software plans retrosynthesis with a Monte Carlo tree search guided by a neural network policy, and it typically finds a solution in less than 10 seconds.
By 2021

Head of Molecular AI

By 2021, he leads the Molecular AI department in Discovery Sciences at AstraZeneca R&D, which develops machine-learning and AI methods for drug design and applies them to the company's small-molecule drug discovery portfolio. He is also an adjunct professor in machine learning and AI for drug design at Chalmers University of Technology.
2024

The REINVENT 4 framework

In 2024, he coauthors "Reinvent 4: Modern AI–driven generative molecule design" in the Journal of Cheminformatics with Hannes Loeffler, Jiazhen He and colleagues from the Molecular AI group in Gothenburg. The open-source framework uses recurrent neural networks and transformer architectures for de novo design, R-group replacement, library design, linker design, scaffold hopping and molecule optimization.
2025

Executive Director and ELLIS Fellow

In 2025, he is Executive Director and Head of Molecular AI in Discovery Sciences within AstraZeneca R&D, and Clarivate names him a Highly Cited Researcher. He is a Fellow of ELLIS, the European Laboratory for Learning and Intelligent Systems.
2026

The ChatInvent agentic system

In 2026, he coauthors a Drug Discovery Today review of ChatInvent, an agentic system based on large language models that AstraZeneca has integrated into its discovery pipeline to aid in molecular design and synthesis planning. The system grows from a single-agent proof of concept into a multi-agent architecture with a graphical user interface.
As of October 2026

Molecular AI at AstraZeneca

He is Executive Director and Head of Molecular AI within Discovery Sciences at AstraZeneca R&D, where he leads machine learning and AI work on molecular design, property prediction and drug discovery. He is an adjunct professor in Data Science and AI at Chalmers University of Technology and a trustee of the Cambridge Crystallographic Data Centre.