Sign in

Tom Diethe

@tomdiethe.com
2.1K followers 415 following 167 posts

Head of the Centre for Artificial Intelligence, DS&AI, Biopharma R&D, AstraZeneca Views are my own

PostsRepliesMedia
Reposted by Tom Diethe
Lichess @lichess.org · 29/09/2026
Bodhana Sivanandan is the youngest WGM in history! 🔥 www.youtube.com/watch?v=EyxG...
0184
Tom Diethe @tomdiethe.com · 08/09/2026
Done!
020
Tom Diethe @tomdiethe.com · 24/08/2026
No cheating. Share the last photo of the sky in your camera roll
000
Tom Diethe @tomdiethe.com · 27/07/2026
Congratulations on both sides!
010
Reposted by Tom Diethe
Clément Canonne @ccanonne.github.io · 02/07/2026
Guest post by John Abowd, @aloni-bologna.bsky.social, Cynthia Dwork, Jae June Lee, @jsarathy.bsky.social, @adamsmith.xyz and Salil Vadhan on the US govt's latest move to ban "noise injection" (at the core of all meaningful privacy protections) from all official statistics: scottaaronson.blog?p=9902
scottaaronson.blog
An American privacy emergency: Guest post from Cynthia Dwork et al.
Scott’s foreword: Cynthia Dwork is Gordon McKay Professor of Computer Science at Harvard, and a pioneer in the fields of differential privacy and algorithmic fairness. On my recent travels to…
0114
Tom Diethe @tomdiethe.com · 11/06/2026
True! For a while I thought random projections were the solution to all problems
010
Tom Diethe @tomdiethe.com · 11/06/2026
Jensen’s inequality in machine learning??
210
Reposted by Tom Diethe
Michael Okun @michael-okun.bsky.social · 10/06/2026
I’ve officially resigned as Associate Editor for Frontiers in Systems Neuroscience. It used to be a reputable journal, but became a case study in how forced automation destroys academic integrity. 👇
5435871430
Reposted by Tom Diethe
ELLIS @ellis.eu · 11/05/2026
🏹 Job alert: Postdoc in LLM Factuality Detection at University of Copenhagen 📍 Copenhagen 🇩🇰 ⏰ Apply by 31 May 🔗 bit.ly/4d6om5H
bit.ly
Postdoc in LLM Factuality Detection
The postdoc’s duties will also to a smaller degree include teaching and supervision.
072
Reposted by Tom Diethe
ELLIS @ellis.eu · 08/05/2026
🏹 Job alert: Postdoctoral Researcher for the ERC Advanced Grant InterText at UKP Lab, TU Darmstadt 📍 Darmstadt 🇩🇪 ⏰ Apply by 31 May 🔗 bit.ly/3Rcg3Ot
bit.ly
Postdoctoral Researcher — ERC Advanced Grant InterText, UKP Lab, TU Darmstadt
023
Tom Diethe @tomdiethe.com · 05/05/2026
Last week's @ebi.embl.org workshop on de novo peptide and miniprotein design was excellent. One theme kept surfacing: data leakage in sequence ML. It's an underappreciated problem — and for peptides especially. Blog post: www.linkedin.com/pulse/why-pr... #MachineLearning #DrugDiscovery
linkedin.com
Why Protein and Peptide Sequence Datasets Leak — and What to Do About It
Tom Diethe | AstraZeneca Last week I was at the EMBL-EBI Industry Programme workshop on From de novo designed mini-proteins and peptides to the clinic — a fantastic event that brought together computa...
000
Tom Diethe @tomdiethe.com · 03/05/2026
Really interesting, thanks for sharing.
060
Reposted by Tom Diethe
ELLIS @ellis.eu · 22/04/2026
🏹 Job alert: Fully funded PhD Position in Test-Time Adaptation and Agentic AI at Aarhus University 📍 Aarhus 🇩🇰 ⏰ Apply by 1 June 🔗 bit.ly/49Tpfgk
phd.tech.au.dk
Readvertisement: PhD Position in Test-Time Adaptation and Agentic AI
0105
Tom Diethe @tomdiethe.com · 12/04/2026
Great article thanks for sharing
010
Tom Diethe @tomdiethe.com · 28/03/2026
Good luck!!
010
Tom Diethe @tomdiethe.com · 24/03/2026
Got sent this by an old friend too. Some primus influence in there too
030
Reposted by Tom Diethe
NeurIPS Europe @neuripseurope.bsky.social · 23/03/2026
Today NeurIPS is announcing our official satellite event in Paris. After responding to the call from Ellis following the success of EurIPS in December, we are pleased to reach a new milestone by joining forces with the NeurIPS organizing committee for the 2026 edition.
18933
Reposted by Tom Diethe
ELLIS @ellis.eu · 23/03/2026
💬 AI for Good Webinar Series 'From Molecules to Models' - today with Alexander Binder 🇩🇪 on 'Explainable AI in cases when you cannot spot the dogs or cats'. 📅 Mon, Mar 23 🕓 17:00-18:00 CET 🔗 Registration: bit.ly/4sP4URk ℹ️ More info: bit.ly/47Qclim
ELLIS AI for Good Webinar Series 'From Molecules to Models' with Alexander Binder on 'Explainable AI in cases when you cannot spot the dogs or cats'.
082
Reposted by Tom Diethe
ELLIS @ellis.eu · 19/03/2026
🏹 Job alert: PhD Position Foundation Models for Industry (FIND) 📍 Delft 🇳🇱 ⏰ Apply by 30 April 🔗 careers.tudelft.nl/job/Delft-PhD-Po…
careers.tudelft.nl
PhD Position Foundation Models for Industry (FIND)
PhD Position Foundation Models for Industry (FIND)
043
Tom Diethe @tomdiethe.com · 14/03/2026
110
Tom Diethe @tomdiethe.com · 14/03/2026
First over the board chess tournament in over 30 years! Games live here (don’t get your hopes up, I have zero expectations!)… www.chess.com/events/briti...
chess.com
British Rapidplay Championship 2026 - Live Games
Watch every game of the British Rapidplay Championship 2026 updated live, move-by-move.
110
Reposted by Tom Diethe
CompSciOxford @compscioxford.bsky.social · 04/03/2026
Recruiting: We're hiring a Postdoctoral Research Associate in Computational Complexity (up to 4 posts) to become a member of a research group with responsibility for carrying out research on the European Research Council Project METACOMP. Details & applications: my.corehr.com/pls/uoxrecru...
A salmon pink graphic with the Department of Computer Science logo at the top left and text reading 'We’re hiring: Postdoctoral Research Associate in Computational Complexity. Full time fixed-term for 2 years starting from 1 September 2026. Grade 7: £39,424 - £47,779 p.a. inclusive of Oxford University weighting. Application Deadline: 12:00 13 March 2026. @compscioxford #CompSciOxford'. There are also lines and dots pointing to the text that look like a computer board.
022
Tom Diethe @tomdiethe.com · 05/03/2026
Any evidence for other types of addiction? Eg gambling?
000
Tom Diethe @tomdiethe.com · 03/03/2026
Reminds me of this! share.google/rj5TFXJumVm4... I’ve wondered this myself too. How much are we just next prediction machines too (where tokens can be abstract concepts such as event streams)? Obviously we also have some kind world model, as well as replay and contractual rollout capabilities
share.google
Google Search
000
Reposted by Tom Diethe
ELLIS @ellis.eu · 02/03/2026
🏹 Job alert: Two PhD positions in Machine Learning: 1. Wallenberg AI, Autonomous Systems & Software Program; 2. KTH Royal Institute of Technology (Sweden) & Nanyang Technological University (Singapore) 📍 Stockholm ⏰ Apply by March 5th 🔗 bit.ly/4cho8Kl
bit.ly
Two PhD positions in Machine Learning | European Laboratory for Learning and Intelligent Systems
054
Reposted by Tom Diethe
ELLIS @ellis.eu · 03/03/2026
🏹 Job alert: PhD and Postdoc Positions in Visual Computing at University of Bonn 📍 Bonn 🇩🇪 🔗 bit.ly/4tWnyb8
bit.ly
Open Positions in ERC Starting Grant project HARMONY ← Learning and Optimisation for Visual Computing
044
Reposted by Tom Diethe
ELLIS @ellis.eu · 02/03/2026
🏹 Job alert: PhD Position in Graph Learning at University of Vienna 📍 Vienna 🇦🇹 ⏰ Apply by March 26th 🔗 bit.ly/4qYURaO
bit.ly
University assistant predoctoral - PhD Position in Graph Learning
University assistant predoctoral - PhD Position in Graph Learning
054
Tom Diethe @tomdiethe.com · 01/03/2026
Looking forward to it!
020
Reposted by Tom Diethe
ELLIS @ellis.eu · 24/02/2026
🏹 Job alert: Scientific Research Associate (Postdoc) in the Field of Agentic AI Systems for Scientific Workflows 📍 Hannover 🇩🇪 📅 Apply by March 25th 🔗 www.uni-hannover.de/en/jobs/id/8358
uni-hannover.de
Scientific Research Associate (Postdoc) in the Field of Agentic AI Systems for Scientific Workflows – Leibniz University Hannover
074
Reposted by Tom Diethe
Katharine Hayhoe @katharinehayhoe.com · 13/02/2026
Bluesky is the new science Twitter, new study by @whysharksmatter.bsky.social and Julia Wester concludes! "Results show that for every reported professional benefit that scientists once gained from Twitter, scientists can now gain that benefit more effectively on Bluesky than on Twitter."
academic.oup.com
Scientists no Longer Find Twitter Professionally Useful, and have Switched to Bluesky
Synopsis. Social media has become widely used by the scientific community for a variety of professional uses, including networking and public outreach. For
9667932211
Reposted by Tom Diethe
CompSciOxford @compscioxford.bsky.social · 13/02/2026
Do you want to study a DPhil (PhD) in the world’s top-ranked computer science department? If so, complete your application before 12:00 Tuesday 3 March 2026. Learn more about the course & how to apply here: www.ox.ac.uk/admissions/g... #CompSciOxford #StudyAtOxford #OxfordUni #ComputerScience
A photograph of a woman in a pale dress with dark hair, pointing with a marker pen at a white board. Over the image there is a grey-green box with blue text reading ‘Are you interested in studying a DPhil in Computer Science at the University of Oxford? Do you want to learn about cutting-edge research while you study in a beautiful, historic setting? Applications are open until 12:00 midday Tuesday 3 March 2026’.
021
Reposted by Tom Diethe
ELLIS @ellis.eu · 13/02/2026
🏹 Job alert: PhD Research Fellow in Machine Learning for Cognitive Neuroscience at the University of Oslo 📍 Oslo 🇳🇴 📅 Apply by March 1st 🔗 bit.ly/3MlLcNy
bit.ly
PhD Research Fellow in Machine Learning for Cognitive Neuroscience (294553) | University of Oslo
Job title: PhD Research Fellow in Machine Learning for Cognitive Neuroscience (294553), Employer: University of Oslo, Deadline: Sunday, March 1, 2026
023
Tom Diethe @tomdiethe.com · 06/02/2026
Very nice!
000
Tom Diethe @tomdiethe.com · 02/02/2026
The next AI revolution isn't bigger models—it's better system architectures. Four emerging shifts: many-to-many collaboration, human-on-the-loop autonomy, richer uncertainty communication, and continual learning. www.linkedin.com/pulse/four-e...
linkedin.com
Four Emerging Shifts in AI System Architecture
The landscape of AI deployment is shifting beneath our feet. While much attention focuses on model capabilities—larger context windows, better reasoning, multimodal understanding—a quieter revolution ...
010
Tom Diethe @tomdiethe.com · 31/01/2026
Welcome 😀 I did this one - but you mean another? go.bsky.app/LBENVXJ
100
Reposted by Tom Diethe
Milot Mirdita @milot.bsky.social · 20/01/2026
My time in @martinsteinegger.bsky.social's group is ending, but I’m staying in Korea to build a lab at Sungkyunkwan University School of Medicine. If you or someone you know is interested in molecular machine learning and open-source bioinformatics, please reach out. I am hiring! mirdita.org
mirdita.org
Mirdita Lab - Laboratory for Computational Biology & Molecular Machine Learning
Mirdita Lab builds scalable bioinformatics methods.
710756
Tom Diethe @tomdiethe.com · 17/01/2026
100
Tom Diethe @tomdiethe.com · 05/01/2026
Never play f6, except when it's checkmate :-). (pls ignore the fact that I missed mate in 1 twice before that!!) www.chess.com/game/live/14...
chess.com
Chess: husamjl1 vs tommydacat
husamjl1 (1078) vs tommydacat (1109). tommydacat won by checkmate in 27 moves. Click to review the game, move by move.
010
Tom Diethe @tomdiethe.com · 26/12/2025
Here!
010
Tom Diethe @tomdiethe.com · 25/12/2025
From my daughter! Merry Christmas all! @lichess.org @agadmator.bsky.social
020
Tom Diethe @tomdiethe.com · 13/12/2025
Tuk a what now?
000
Tom Diethe @tomdiethe.com · 12/12/2025
Fun visiting @harvardmed.bsky.social earlier today!
010
Tom Diethe @tomdiethe.com · 04/12/2025
Nice talk today!
010
Tom Diethe @tomdiethe.com · 04/12/2025
The problem (in general) with Norwegian salmon is that it’s reared in the calm waters of the fjords. Leads to fatty fish (not enough exercise) and a lot of disease. Scottish salmon (in general) is better, as the farms are in rougher waters, solving both of these issues
030
Tom Diethe @tomdiethe.com · 11/10/2025
960 solved
000
Tom Diethe @tomdiethe.com · 11/10/2025
Is this a version of Cauchy-Schwartz?
100
Tom Diethe @tomdiethe.com · 11/10/2025
Yep, mainly phone calls
000
Reposted by Tom Diethe
Neil Lawrence @lawrennd.bsky.social · 07/10/2025
Volume 298 proceedings.mlr.press/v298/ Proceedings of Machine Learning for Healthcare 2025 Is now available on PMLR
proceedings.mlr.press
Proceedings of Machine Learning Research
Proceedings of the 10th Machine Learning for Healthcare Conference Held in Mayo Clinic, Rochester, MN, USA on 15-16 August 2025 Published as Volume 298 by the Proceedings of Machine Learning Research ...
042
Reposted by Tom Diethe
ELLIS @ellis.eu · 07/10/2025
How can we make probabilistic ML fast, generalizable, and ready for the real world? 🤔 Join us for the ELLIS UnConference Workshop on Amortized ProbML on Dec 2 in Copenhagen 🇩🇰 co-located with #EurIPS 📣 Call for contributions open until Oct 10 👉 Details: eurips.cc/ellis
141
Reposted by Tom Diethe
Patrick Kidger @patrickkidger.bsky.social · 07/10/2025
🚀 New talk! "Automated ML-guided lead optimization: surpassing human-level performance at protein engineering" ▶️ www.youtube.com/watch?v=mEhB... ✨🧪 This was a talk I gave at the recent AIxBIO conference in Cambridge UK. A 10-minute pitch for what we do at Cradle!
youtube.com
Automated ML-guided lead optimization: surpassing human-level performance at protein engineering
YouTube video by Patrick Kidger
0144