Jan Witowski (@janwitowski) 's Twitter Profile
Jan Witowski

@janwitowski

CEO & co-founder at @Ataraxis_AI. MD, PhD. Personalizing cancer treatment. Previously @nyugrossman @harvardmed

ID: 56081214

linkhttp://ataraxis.ai calendar_today12-07-2009 12:24:18

869 Tweet

1,1K Followers

1,1K Following

Eric Topol (@erictopol) 's Twitter Profile Photo

False positives and unnecessary biopsies frequently arise from mammography screening for breast cancer. A new study using #AI of MRI scans shows the potential for much improvement science.org/doi/10.1126/sc… Science Translational Medicine Jan Witowski Krzysztof Geras New York University NYU Grossman School of Medicine and colleagues

False positives and unnecessary biopsies frequently arise from mammography screening for breast cancer. A new study using #AI of MRI scans shows the potential for much improvement science.org/doi/10.1126/sc… <a href="/ScienceTM/">Science Translational Medicine</a> <a href="/JanWitowski/">Jan Witowski</a> <a href="/kjgeras/">Krzysztof Geras</a>
<a href="/nyuniversity/">New York University</a> <a href="/nyugrossman/">NYU Grossman School of Medicine</a> and colleagues
NVIDIA Healthcare (@nvidiahealth) 's Twitter Profile Photo

A study led by Jan Witowski and Krzysztof Geras employed #DNNs to develop an #AI model for detecting #breastcancer in dynamic contrast-enhanced MRI scans. The model matches the accuracy of experts and could reduce unnecessary biopsies. nvda.ws/3EgGYjR

Krzysztof Geras (@kjgeras) 's Twitter Profile Photo

A paper summarizing a challenge that our team from NYU won. It is not easy to translate results from challenges to solutions that will be useful in practice, but they are an important reality check! I would love to see even more challenges on medical AI and their analyses.

Yann LeCun (@ylecun) 's Twitter Profile Photo

SAM: Segment Anything Model from FAIR. Foundation model for image segmentation. Demo: segment-anything.com/demo Blog: ai.facebook.com/blog/segment-a… Paper: ai.facebook.com/research/publi… Code: github.com/facebookresear… Dataset: SA-1B , 11 million image, 1 billion masks ai.facebook.com/datasets/segme…

Yann LeCun (@ylecun) 's Twitter Profile Photo

I-JEPA: Efficient method for Self-Supervised Learning of image features. No need for data augmentation, just masking. Joint embedding predictive architecture, not generative. And it's open source, of course. Blog: ai.facebook.com/blog/yann-lecu… Paper: arxiv.org/abs/2301.08243 Code &

Jan Witowski (@janwitowski) 's Twitter Profile Photo

It’s time to use AI to help cancer patients get truly personalized treatment! To solve this, I founded Ataraxis AI with Krzysztof Geras. Ataraxis is an AI precision medicine company building a new era of tools for better treatment selection, starting with Ataraxis Breast:

Ataraxis AI (@ataraxis_ai) 's Twitter Profile Photo

ā€œIf you don’t benefit from chemotherapy, you want to avoid it.ā€ Our CEO Jan Witowski shared with Forbes' Alex Knapp how Ataraxis Breast enables providers to recommend more precise, effective treatments—helping patients avoid unnecessary chemotherapy—and more.

Jan Witowski (@janwitowski) 's Twitter Profile Photo

We are hiring at Ataraxis! Jobs: jobs.ashbyhq.com/ataraxis-ai The mission: apply advanced AI to one of the hardest problems in healthcare—optimizing treatment and predicting patient outcomes Looking for problem-solvers across all roles who want to build something meaningful

We are hiring at Ataraxis! 
Jobs: jobs.ashbyhq.com/ataraxis-ai

The mission: apply advanced AI to one of the hardest problems in healthcare—optimizing treatment and predicting patient outcomes

Looking for problem-solvers across all roles who want to build something meaningful
Ataraxis AI (@ataraxis_ai) 's Twitter Profile Photo

We are coming to NeurIPS! Joe Cappadona will present results of our novel approach to training foundation models for digital pathology. This research goes beyond simply scaling model/data size, and led us to develop Kestrel, the model behind Ataraxis Breast. Come see us!

We are coming to NeurIPS! <a href="/joecappadona/">Joe Cappadona</a> will present results of our novel approach to training foundation models for digital pathology.

This research goes beyond simply scaling model/data size, and led us to develop Kestrel, the model behind Ataraxis Breast. Come see us!
Jan Witowski (@janwitowski) 's Twitter Profile Photo

I will be at the San Antonio Breast Cancer Symposium #SABCS24 next week sharing amazing clinical validation results of Ataraxis Breast--the first AI-native prognostic/predictive test for breast cancer patients

Jan Witowski (@janwitowski) 's Twitter Profile Photo

Ataraxis has a really unique combination of world-class AI researchers and medical/commercial leaders who brought to market new products used by millions of patients every year. We have only a few more open roles left:

Krzysztof Geras (@kjgeras) 's Twitter Profile Photo

Big personal update! As of 3/1, I’ve joined Ataraxis AI — a startup I co-founded in 2022 — full-time as Chief Scientific Officer. We recently raised a $20.4M Series A, and are now building a frontier AI lab to push the boundaries of personalized oncology treatment. 1/

Krzysztof Geras (@kjgeras) 's Twitter Profile Photo

We are looking for a biostatistician to join us: jobs.ashbyhq.com/ataraxis-ai/a0… This an amazing position for someone with strong skills in biostatistics, interested in optimizing cancer treatment with AI. Please help me spread the word!

Ataraxis AI (@ataraxis_ai) 's Twitter Profile Photo

We are at #ASCO25 this weekend! Jan Witowski will be presenting the results from our Ataraxis Breast clinical validation study--stop by to learn how our multi-modal AI and foundation models help personalize breast cancer treatment.

We are at #ASCO25 this weekend! <a href="/JanWitowski/">Jan Witowski</a> will be presenting the results from our Ataraxis Breast clinical validation study--stop by to learn how our multi-modal AI and foundation models help personalize breast cancer treatment.
Jan Witowski (@janwitowski) 's Twitter Profile Photo

Grok 4 (Heavy and base) has been a disappointment--mostly the product, not model. Somehow it fails to maintain context even after 2-3 messages, fails to work consistently with files, randomly decides to searches the web even with the option disabled.. o3(-pro) still wins so far.