ReactWise

ReactWise x AI4Green: Bridging the gap between lab records and optimization. The "connected lab" is a common goal in R&D, but the reality is often fragmented. When experimental data lives in a siloed ELN, it can’t be easily used to drive real-time decision-making, and bridging that gap manually is often slow and error-prone. Our latest integration with AI4Green - a sustainability-focused ELN developed by Professor Jonathan Hirst’s group at the University of Nottingham - addresses this directly. A two-way API integration connects laboratory records to optimization workflows. Historical ELN data flows into ReactWise to support easy use of transfer learning methods, such as MemoryBO, warm-starting new campaigns with relevant prior knowledge. In return, data generated within ReactWise is sent back to the ELN, maintaining a complete and continuous experimental record. The result is less manual data handling, stronger data continuity, and a more connected digital lab environment where optimization and record-keeping reinforce each other. This is a step toward the kind of staged, practical infrastructure we believe makes autonomous experimentation achievable - starting with structured data capture and building from there. We’re excited to partner on this with teams like AI4Green who share our commitment to a more connected, sustainable, and data-driven laboratory. Read the full case study on our website - link in the comments.

Interested in the intersection between flow chemistry and machine learning? The two should work together perfectly; in an ideal world, you can set up the flow system, connect it to an inline PAT for online analytics, and feed the results into optimization software that can control flow rates. Once this has been left to run, it should result in an optimized system, with no further intervention needed. Neat case studies often show promising results, but the challenges of repeating this in real-world optimizations mean it is never simple. Our CEO, Alexander Pomberger, is heading to Flow Chemistry Europe in Málaga this April to discuss exactly that: Bridging the gap between lab-scale optimization and real-world industrial impact. Get the pre-read in the comments. Otherwise, come along to Málaga and listen in.

It was great to see our HTE chemist, Kaiman Cheung, back at his alma mater last week, speaking on a careers panel at the University of Bristol. At ReactWise, Kaiman focuses on the generation of large experimental high-quality datasets that feed our zero-shot prediction models, including HTE plate design, execution, and quantification. The panel covered a lot of ground, and Kaiman's contributions centred on AI's role in chemistry. AI won't “solve chemistry”, but will become part of how chemistry is done. The question is in getting the balance right - where can different algorithms and models actually support chemists - so we have to be honest about the trade-offs: AI has a real footprint, but so does running unnecessary reactions, wasting solvents, and scaling bad processes. He also spoke about what it is like to work at ReactWise. When generating large datasets, the value added lies in choosing the right experiments, conditions, and structure to make the result useful for our clients. This range of responsibilities and ownership is just part of how a startup works. It's always good to see our team sharing insights into working at a chemistry-first software startup.

Pfizer x ReactWise: Multi-Objective Hydrogenation Process Optimization We're proud to share our latest case study, developed in collaboration with our clients at Pfizer. Together, we set out to tackle a complex process optimisation challenge: finding the ideal temperature profile for a hydrogenation process to maximize yield and minimize impurity. The Pfizer team used a multi-objective Bayesian Optimization workflow within the ReactWise platform. This enabled effective initial exploration of the broad design space - hold temperatures, hold times, and ramp rates - followed by rapid convergence on the ideal conditions. The outcome was achieving targets of 97.5% yield and sub-1% impurity while also uncovering a critical interaction between temperature profile and process performance. It’s incredibly rewarding to share examples of data-driven optimization solving real-world industrial challenges. Stay tuned as we share more success stories in the coming weeks! Find the full case study on our website - link in the comments.

We are hiring a front-end engineer intern to join our team in Cambridge. We’re a team of chemists, machine learning engineers, and software developers, backed by Y Combinator & Innovate UK - working with several of the largest pharma companies. We are looking for someone to build intuitive, responsive user interfaces with React and TypeScript. This is a great opportunity to grow your skills in a supportive, high-impact environment, with lots of ownership and mentorship.   The details: Role type: 10-week internship Dates: Flexible, between early spring and late autumn Location: In-person/hybrid; Cambridge, UK Profile: Current undergraduate, or recent grad Find the detailed role description and application link on our website - link in the comments.