Shrey Jain
@shreydiwjain
AI Product Lead @Microsoft. Previously Special Projects @MSFTResearch, Eng Sci @UofT, CSAIL @MIT. Views are my own.
I'm happy to announce the launch of the Healthcare Agent Orchestrator, a project I led at Microsoft with an incredible team! Watch @satyanadella's keynote presentation highlight our incredible collaboration with @StanfordHealth.
Stanford Medicine is orchestrating multiple agents to connect medical data, clinical trials, and more - all with the goal of delivering more effective cancer care. Incredible to see in action.
Excited for MedHELM to be out! This is an open-source clinician-validated framework built to evaluate medical AI on real-world tasks, offering a more meaningful alternative to traditional exam benchmarks and helping the community build safer, more reliable models for medical…
🧵 🩺 LLMs score ~99% on medical licensing exams, but are they ready for real medical deployment? Our new research reveals major gaps between test performance and clinical readiness. Introducing MedHELM: Holistic Evaluation of Large Language Models for Medical Applications👇
Biomarkers can be the difference between life and death, but only ~half of U.S. lung cancer patients are tested for biomarkers for which effective targeted drugs are readily available. This new @nature Medicine paper from @EliLillyandCo shows how fine-tuning GigaPath (from…
Humans are constantly operating under uncertainty, and one way in which we increase confidence in our decision making is by asking questions! When should a model ask a question if it’s uncertain? This is critical to understand in healthcare where the value of more information…
We transform these cases into the Sequential Diagnosis Benchmark, a more realistic way to assess AI. Models start with minimal information about a case, and need to ask questions, order tests, interpret results, and refine their hypotheses along the way - like clinical practice.
Harnessing AI to help with difficult cancer decision problems: A pleasure to join @shreydiwjain for a fireside conversation on pathways forward for translating AI advances into tangible tools for enhancing clinical care @MSFTResearch @mattlungrenMD @Microsoft @StanfordHealth
Had a great conversation with @erichorvitz, Microsoft’s Chief Scientific Officer, on how AI is transforming cancer care. We explored how AI can enhance tumor boards and bring more personalized, data-driven care to patients -now made possible through Microsoft’s Healthcare Agent…
We have only seen glimpses of the impact @RichardSSutton and Andrew Barto’s work on reinforcement learning will have on humanity. Huge congratulations to both of them being awarded the @TheOfficialACM Turing Award last night!
It’s been a pleasure collaborating with @shreydiwjain & colleagues at @Microsoft’s Health Care & Life Sciences and @MSFTResearch on directions ahead for leveraging AI advances to help with cancer care. @StanfordHealth @ARPA_H
Tumor board meetings are where some of the most critical decisions in cancer care are made. Here is a preview of a conversation between @erichorvitz and I on why these meetings are so vital, informing the first of many use cases for the Healthcare Agent Orchestrator.