DMTA Innovation at Takeda | Atinary Webinar Series Ep03
Webinar Recap: Takeda’s journey to a Self-Driving chemistry workflow with AI and automation, cutting a 1–2 month manual process to 1 week while reducing experiments and removing bottlenecks.
📍 Online Webinar | June 24, 2025
🎙️ Speakers:
Dr. Adrian Ramirez Galilea (Associate Director of Automation & HTE, Takeda)
Dr. Hermann Tribukait (Co-Founder and CEO, Atinary Technologies)
Dr. Alexander Waker (Katalyst D2D Product Coordinator, ACD/Labs)
Transforming R&D from months to days. That was the journey we explored in our June webinar with Takeda and ACD/Labs. Read on to see how Takeda’s HTE & Automation team built a self-driving, closed-loop system for synthetic molecule process development, and how AI and automation are accelerating discovery in real-world chemistry workflows.
From Manual Efforts to a Dedicated HTE & Automation Team
Three years ago, Takeda’s high throughput experimentation (HTE) & Automation group in Cambridge, US, began as a small team of isolated subject matter experts (SMEs) and instruments. Back then, their work relied heavily on labor-intensive, manual processes, which created significant bottlenecks:

- Inefficient use of resources
- Time-consuming manual work
- Duplicated documentation
- Lost knowledge
Tackling these pain points became the team’s mission, driving the evolution toward an integrated, efficient, and data-driven approach to process development. This shift required not only robotics and automation, but also an AI layer to design the next experiments, interpret the resulting rich data, and guide the team efficiently through complex chemical spaces.
“Prior to this integration, we had to transform the data into different formats, running also manual composition calculations. It could take up to 4 hours to have everything ready to start the AI-guided run in the Unchained system, being now a matter of minutes.”
Dr. Adrian Ramirez Galilea, Associate Director, Automation & HTE
Building the Self-Driving Data-Rich HTE Approach
To enable seamless, AI-guided closed-loop optimization, Adrian and his team combined hardware and software to minimize or remove bottlenecks. Their tech stack includes:
- Unchained Labs robotics for automated consumables and liquid handling
- ACD/Labs’ Katalyst D2DⓇ for automated data collection & visualization
- Dotmatics for ELN and data management
- Atinary SDLabs™ for AI-driven experiment design & optimization
“We need a smart combination of hardware and software, where all the bottlenecks are removed
Dr. Adrian Ramirez Galilea, Associate Director, Automation & HTE
or minimized (i.e. analytical data processing and software / hardware connectivity.”
This approach delivered tangible results:
- Reliable data for scaling from 100 μL to 1 mL with fewer robotic errors
- Overcome the low throughput by AI-guided optimization across generations to reach desired yield
- Iterative sampling without disturbing reactions (4 × 50 μL, multiple timepoints)
- Fully automated workflows leveraging mobile robotics to operate 24/7

Case Study: Optimizing a Buchwald-Hartwig Reaction
“The optimization was done with our Self-Driving approach in about 1 week,
Dr. Adrian Ramirez Galilea, Associate Director, Automation & HTE
in comparison to the 1-2 months that would have taken the manual external approach.”
A highlight of the webinar was a case study on optimizing a Buchwald-Hartwig reaction:
- Parameter space:
6 bases × 5 equivalents × 6 Pd pre-catalysts × 4 catalyst equivalents × 6 solvents = 4,320 possible combinations. - Traditional timeline: 1 to 2 months, manually.
- With Atinary SDLabs + automation: 1 week.
5 parameters | 4-6 conditions per parameter | 4320 possible combinations |
With Atinary’s SDLabs analytics, the process development team could gain clear insights into how conditions affect performance across experiments. Contour plots highlighted clusters of robust, scalable regions versus isolated optima where small shifts could cause sharp yield losses. Parallel coordinate plots provided intuitive visualizations, highlighting parameter contributions to maximizing yield and peak area.
The Outcome: Beyond the Lab
Takeda’s journey illustrates how AI-driven workflows simplify complex R&D challenges. By combining robotics, digital automation, and intuitive AI, process development can move from months to days. Benefit include:
- Eliminating repetitive manual work
- Reducing bottlenecks
- Preventing duplicated efforts
Whether starting your AI journey or already leveraging it, the SDLabs Platform supports workflows ranging from AI-guided DoE recommendations to fully closed-loop autonomous optimization.
The future of science is self-driving, AI-enabled labs, and it’s happening today.
🎥 Watch the full webinar here: https://www.youtube.com/watch?v=KeCtyDAKiZ0
Resources
- Link to Press Release: Collaboration with Takeda
- Link to Use-Case with Takeda on Leveraging HTE to optimize yield in deprotection reaction
- Link to C&EN Webinar: Leveraging HTE & Automation for Synthetic Molecule Process Development (sponsored by Unchained Labs)
- More about Takeda Pharmaceuticals