
Applied AI Leader, ML/Deep Learning, Genomics & Precision Medicine. Formerly @AstraZeneca, @Cambridge_Uni and @emblebi | https://xitter.cf/t.co/PzKNkB4Jvy
Cambridge, England
Joined January 2017
- Tweets232
- Following446
- Followers286
- Likes628
What really sits behind an #AI #agent?
The model is only one part of the system. Tools, memory, guardrails, orchestration, etc... are what help turn raw model capability into something useful.
A few thoughts on the AI agent #harness:
dvitsios.org/2026/07/31/agen…
🤖 Made with AI
#AI has made starting work almost free.
But...we are producing far more work than we can realistically supervise.
The next frontier in AI may not be intelligence.
It may be #reliability.
I explore this idea and a concept I call #cognitive gates:
👉 dvitsios.org/2026/03/11/the-…
We’re hiring a Senior AI Scientist at Quotient Therapeutics in Cambridge, UK!
Join our mission to apply #AI & #ML in #LifeSciences and tackle big challenges in #SomaticGenomics.
🚀 Apply now:
linkedin.com/jobs/view/41940…
#AI #Hiring #SomaticGenomics
Glad to be part of this collaboration, featuring #mantisml 🚀 for prediction of genes associated with neurodevelopmental disorders
📣 New today!
📄Genome-wide prediction of dominant and recessive neurodevelopmental disorder-associated genes
🧑🤝🧑@RyanDhindsa @awzoghbi & colleagues
cell.com/ajhg/abstract/S0002…
- Breakthrough in quantum computing – 30K readers
- Incremental smartphone update – 🚀 200K+ readers
Spot the mistake
I’m grateful yet again to some of the books that kept me company several months during this year 🙏
Whenever I read a book, I still feel like a kid, opening a new ‘door’, over-excited about what I’m going to discover (or create) in this new world..
Happy Christmas!🎄
Delighted to see our #MILTON work being presented at #ASHG24 by @manik_garg_1 in a packed room! ⭐️
"Disease prediction with multi-omics and biomarkers empowers case-control genetic discoveries in UK Biobank."
Read more in our #NatureGenetics paper!
nature.com/articles/s41588-0…
All eyes on JARVIS 2.0 ⭐️
Anna Maria Tsakiroglou presenting:
"JARVIS 2.0: A non-coding variant effect predictor based on 500,000 UK Biobank WGS samples that surpasses existing benchmarks for prioritisation of common variants."
#ASHG24 #ReviewersChoice #Award
Dimitrios Vitsios retweeted
Excited to share our latest preprint where we describe how diverse ancestral representation in sequencing databases improves genetic intolerance estimates!
biorxiv.org/content/10.1101/…
Dimitrios Vitsios retweeted
@StankovicStasa"Multi-omics framework discovers novel biomarker and genetic signatures for women's reproductive health" poster presentation 7042F
Xiaoyin Li "Fast and efficient toolkit for rare variants association testing" poster presentation 4061F
Dimitrios Vitsios retweeted
@dimvitsios: "BinBVS: Bayesian variable selection for classification of binary features, applied to multi-genic rare variant data from Phenome-wide association studies in UK Biobank" poster presentation 4026F
Dimitrios Vitsios retweeted
Anna Maria Tsakiroglou: "JARVIS 2.0: A non-coding variant effect predictor based on 500,000 UK Biobank WGS samples that surpasses existing benchmarks for prioritisation of common variants." Reviewers’ choice poster 1146T
Dimitrios Vitsios retweeted
@SlavePetrovski, myself and @dimvitsios: “Inside AstraZeneca’s Genomics Initiative: genomics-driven drug discovery.” Wed, Nov 06 12:00pm – 12:30pm CoLab Session Theater 3
Dimitrios Vitsios retweeted
@manik_garg_1 : "MILTON: Disease prediction with multi-omics and biomarkers empowers case-control genetic discoveries in UK Biobank." Wed, Nov 06 10:15am - 11:45am Machine Learning and AI Applications in Human Genetics session
Dimitrios Vitsios retweeted
Thrilled to be in Denver with 14 colleagues from @AstraZeneca’s Centre for Genomics Research, ready for 5 days of human genomics #ASHG24
The perfect chance to learn about how genomics is transforming drug discovery!
Here’s a thread of our presentations - see you there 😀
Dimitrios Vitsios retweeted
I am thrilled to be presenting #MILTON at #ASHG24 ⭐. Come and learn more about this machine learning tool on 6th Nov, 10:45 am-11:00 am in the Four Seasons Ballroom 2&3 at Colorado Convention Center. See you soon!
Early disease detection = better patient outcomes. Machine learning tool MILTON can predict disease well before symptoms present. Out yesterday in @NatureGenet, using @uk_biobank data. Congratulations CGR team, esp @manik_garg_1
@marcinkarpinski @SlavePetrovski and @dimvitsios!
Understanding the Encoder-Decoder Architecture of Transformers…one word at a time: 📚
dvitsios.org/2024/09/22/unde…













