Edvin Fako
@efssh
Followers
318
Following
507
Media
23
Statuses
141
Computational Chemistry and Machine Learning @SchwallerGroup @EPFL | @BASF | @TheorHetCatICIQ | scientific Illustrations when time allows
Lucerne, Switzerland
Joined February 2019
🔥 AutoAdsorbate 🔥 has been the workhorse of high throughput computation in heterogeneous catalysis modeling and ML @BASF for the previous few years. 🧵 Now it is available open source! #MatSci #CompChem Preprint: https://t.co/KzkAzkKx0D Code: https://t.co/p0nfGeDNzc
1
20
59
Exciting postdoc opportunity in the @SchwallerGroup at EPFL! We're hiring a postdoc to advance ML-driven synthesis planning after Zlatko Joncev’s successful exit to co-found B-12 (YC '25) 🚀 Work on: - LLMs for strategic synthesis planning - Chemical reasoning at scale -
arxiv.org
While automated chemical tools excel at specific tasks, they have struggled to capture the strategic thinking that characterizes expert chemical reasoning. Here we demonstrate that large language...
0
17
38
Which foundation MLIP is best suited for your application? The MLIPX framework can help you answering this question. Checkout the Paper https://t.co/YYOfLpchqP and https://t.co/ZFh037oars
#MLIPX
github.com
Machine-Learned Interatomic Potential eXploration (mlipx) is designed at BASF for evaluating machine-learned interatomic potentials (MLIPs). It offers a growing set of evaluation methods alongside ...
1
2
6
Check out the fantastic entries, adsKRK and DynoAgent, from the @SchwallerGroup's LLMs for Materials and Chemistry Applications Hackathon team! From heterogeneous catalysis to proteins - MD simulations made simple and explainable 🚀 @efssh @XuanVuNguyen18 @6ojaHa Ryo, Salomé,
You don’t like molecular dynamics? We get it. That’s why at this year’s LLM hackathon for Chemistry and Materials Science, we built not one, but ✨two✨ AI agents for molecular dynamics 👇
1
3
18
High-dimensional linear mappings, or linear layers, dominate both the parameter count and inference cost in most deep learning models. We propose a general-purpose drop-in replacement with a substantially better capacity - inference cost ratio. Check it out!🧵
0
0
4
🚨 PhD & Postdoc positions in Computational Chemistry 🚨 Funded by SNSF + ERC Starting Grant in my group @unifrChemistry (Fribourg 🇨🇭). Focus: ML × DFT (method development & applications). Please RT 🙏 📄 Info & apply:
3
58
149
Efficient parametrization of complex surface structures with autoadsorbate: https://t.co/5IX4BpcCVu
github.com
Chemical intuition for surface science in a package. - basf/autoadsorbate
0
0
2
Psi-k @Psik_Network was intense & inspiring! Great to meet new & old faces (finally in 3D, not just 2D 😅)! Loved sharing our work on reactive interfaces on the big stage. Hope our ideas help make sampling complex interfaces a bit easier. Soon more to come from @SchwallerGroup!
1
5
30
There is a lot of passion, knowledge, and experience that @ZJoncev @drecmb bring, and it is well reflected in the video! Congrats on the launch!
b-12 ( https://t.co/sUMDUAK8TT) is building AI agents that help chemists design experiments and automatically run them on lab robots. They turn months of manual chemistry work into minutes of automated execution. Congrats on the launch @drecmb & @ZJoncev! https://t.co/waQzlnrS7z
0
0
1
Here is the notebook I used to prepare the frames: https://t.co/l9U0PQacZa After that is just a matter of making things look pretty, with #blenderforscience for example... Here's a tutorial on that: https://t.co/pSTFfd6nrj Enjoy 🫠
0
1
6
The idea of what was suposed to be a slide... Surface atoms: easy to see, hard to define: - Z-cutoff? Too blunt. - Coordination? Needs finesse. - CatKit/Pymatgen? Powerful but opaque. Surface roughness & reconstructions don't care. Neither does 🔥AutoAdsorbate🔥!
0
0
0
Started with some slides for AutoAdsorbate-ended up with an animated deep dive into surface site detection. It works on any surface, no bulk symmetry needed. Here’s what this one-liner really does: s=Surface(slab211) code: https://t.co/p0nfGeDNzc paper: https://t.co/3ZdXPgqo3Q
3
13
52
Learn how to render multiple molecules from a single multi-frame .xyz file in Blender! 🧪 Hands-on session recording from the AIChemist School: bridging explainable AI and chemistry. @SchwallerGroup Watch here: https://t.co/RjoTcnIdrU More info: https://t.co/xDEjiKtznB
github.com
Useful tips and tricks for chemists in a hurry, trying to get started with Blender 3D. - schwallergroup/blender-aichemist
2
11
38
Thrilled to present our new work on Reinforcement Learning for 3D Crystal Diffusion Generation in #AI4Mat at #ICLR2025 ! We show that RL can optimize diffusion models for goal-directed crystal generation. Thanks to @pschwllr and @SchwallerGroup ! https://t.co/zwAYYd5PyA
openreview.net
Recent advances in diffusion models have enabled increasing capabilities for inverse materials design. The key capability is achieving tailored design towards desired property profiles, with wide...
1
5
13
Hello from Singapore 🇸🇬! Thrilled to be at #ICLR2025 presenting our work on fragment-based drug discovery 🧩. We go beyond virtual screening with a generative, structure-aware approach. 📃 https://t.co/rH6SkSCPIY 🔗 https://t.co/H8BTPGLZp8 A thread 🧵👇
github.com
Structure-based fragment identification in latent space - rneeser/LatentFrag
1
19
50
9/🧵 Finally! Massive kudos to all of my friends and former colleagues @BASF that helped shape AutoAdsorbate into what it is today. Particularly @SandipDeScience! Special nod to the #MACEMP developers, the universal MLIP mace-mp0 is 👌 Code:
github.com
Chemical intuition for surface science in a package. - basf/autoadsorbate
0
0
3
8/🧵 With this kind of collective view of the energetics over different surfaces, we (and 💻) can really appreciate the origin of their divergent behavior... (More details in the next days)
1
0
1
7/🧵 Let's unpack this dense info. Each signal in this plot comes from a reasonable initial structure that was relaxed to a local minima using MACE-mpO (energy reference - most stable structure for each empirical formula). Comparing two PSE diagonal neighbors Cu vs Pd (211):
2
0
1
6/🧵 Why should anyone care? Cu is *the* metal for het. cat. CO2 reduction. It's safe to say that it's a well studied system. It turns out that the energetics are more complex than anticipated! We relaxed ALL intermediates (~100k) to get the FULL picture:
1
0
2
5/🧵 What's new? We can identify all sites on 🔶️ANY SURFACE🔶️ structure without any additional info. We can use 🟠*SMILES🟠 to generate 🟧ANY MOLECULAR FRAGMENT🟧 (molecule or intermediate) and sample their conformational space. No limitaions on size and complexity.
1
0
1
4/🧵 Meet a new approach implemented in 🔥AutoAdsorbate 🔥 It's designed to GENERATE accurate, chemically meaningful configurations of any molecule on any surface no human guesswork required. The ingredients? Simple heuristics. The result? Let's take a look 🔬⚡️
1
0
1