Salt AI

We're #hiring a new Solutions Engineer in United States. Apply today or share this post with your network.

Top 5 Takeaways from AIDDD 2025 1. The industry is aligned on moving from point solutions to platforms. As Sagar Jain emphasized in his opening, there’s broad consensus across biopharma that the future lies in AI platforms, not isolated point solutions. The vision is shared: integrated data, modular models, and unified scientific workflows.  2. The real traction is in Data → Models → Workflows synergy. Across pharma, biotech, and AI-native companies, the strongest resonance was around integration, not individual tools. This came through clearly in my conversation with @Hao Zheng, Lilly — where the concept of an AI marketplace emerged as a practical way to unify internal and external models, data assets, and workflows into one governed ecosystem. Even teams with strong internal ML builds are actively looking outward for solutions that can tie everything together. 3. Some R&D stages are still underserved — especially pre-IND ADME/Tox. While AI workflows exist for much of the lifecycle, gaps remain in early toxicology and pre-IND decisioning. In my discussion with @Claudette Fuller, Genmab, we dug into how ADME/Tox and early safety remain fragmented and how much opportunity there is for structured, automated, context-rich workflows in these stages. 4. Contextual AI will define the next decade of biopharma workflows. The next wave of AI in drug development won’t be driven by bigger models — it will be driven by context. Teams are realizing that true scientific decision support requires systems that understand the rationale behind experiments, the history of prior work, domain constraints, business rules, and the “why” behind recommendations. This came through clearly in my discussion with Gangadhar Sunkara, Novartis, who emphasized that empowering scientists with the data they generate requires more than predictive power — it requires an AI that can interpret that data in the right biological, clinical, and operational context. 5. And lastly — the people make the event. AIDDD continues to be a special gathering of scientists, engineers, product thinkers, and platform builders. Amazing to reconnect with familiar faces and meet new people pushing the field forward. The energy around collaboration in this space grows every year.

We're #hiring a new Full Stack Engineer in United States. Apply today or share this post with your network.

Biopharma strategy development typically requires months of consulting coordination. We modeled a high-performing consulting team using specialized AI agents. Each handles distinct expertise: strategic analysis, regulatory considerations, trial design, implementation planning. This type of draft workflow addresses adaptive trial strategy using basket trial approaches for oncology therapeutics. Traditional consulting runs sequentially. This approach processes in parallel. Strategic insights emerge faster, regulatory considerations integrate immediately, trial design iterations happen in real-time. Companies that iterate R&D strategies quickly based on emerging data will outpace those locked into quarterly consulting cycles. The broader implication: any complex problem requiring coordinated expert analysis can use this model swarm approach. Visual AI workflows remove the coordination bottleneck. Watch the walkthrough: https://lnkd.in/gwtFMSCT #visualAI #aiworkflows #biopharma #modelswarm

My kids will live to 150. Not because of science fiction, but because telomere research is unlocking mechanisms that slow aging. Telomere length in early life predicts lifespan, and the shortening rate correlates more strongly with longevity than initial length. This reminds me of Rockefeller University's pattern. Founded in 1901 after John Rockefeller's grandson died from scarlet fever, it produced discoveries that changed medicine: DNA as the basis of heredity, blood groups, the first blood bank, Peyton Rous proving cancer can be caused by a virus. What made Rockefeller remarkable was methodology. Small teams of exceptional scientists solving fundamental problems. Karl Landsteiner discovered blood groups in 1901. Rous proved viral cancer in 1910, though it took 56 years for his Nobel. The parallel to the Ellison Medical Institute is striking. Dr. Agus built a vertically integrated research center applying AI to cancer therapeutics. The approach mirrors Rockefeller: multidisciplinary teams, technology-enabled research, focus on disease mechanisms. At Salt AI, we're accelerating this model. When computational biologists can adjust protein folding models directly through visual workflows, research cycles compress from months to weeks. Obviously, I don't know exactly how long my kids will live, but it will be longer than me, and their kids longer than them. The infrastructure decisions we are making today will accelerate biological research and determine that timeline. #telomeres #longevity #bioresearch

How does Salt fit into your AI stack? Salt AI isn’t an MLOps platform. It builds on MLOps. While MLOps runs models efficiently, Salt connects models, data, and tools so they can reason and act together. It’s an orchestration layer for intelligence on top of inference. Salt SuperIntelligence enables context-aware systems where models call tools, tools return data, and everything adapts in real time. The result: collaborative AI that moves beyond isolated tasks to coordinated action across your stack. Learn more: https://www.salt.ai/

Security audits and compliance reviews slowing down your research initiatives? We get it. In life sciences, the bar for data security is exceptionally high—and it should be. That's exactly why we pursued (and achieved) SOC 2 Type 1 certification. Independent auditor Sensiba LLP validated our security controls with zero exceptions, so you can focus on breakthroughs instead of compliance concerns. Your research deserves infrastructure built on trust. We're delivering it. See what went into this certification and what it means for your team: https://lnkd.in/g-tJn5S2 #LifeSciences #Biopharma #AIInfrastructure #Compliance #ResearchInnovation #HealthcareTechnology #SOC2

Today we welcome four accomplished leaders from BCG, Doctor Evidence, Concert AI, and Thermo Fisher as we scale Salt AI’s adaptive pipeline across life sciences—following last month’s $10M raise. Details → https://lnkd.in/g9CxWc6m See the platform live at AIDDD Summit (Nov 19, Boston) with demos from @Toby Sayre and @Dan Sheikh. Aber WhitcombJim BenedettoCharles BasilRyan WangMary Grace GraczykNate BeyorBob BattistaTyson Butler

Next week, Salt AI’s Toby Sayre and Dan Sheikh will be in Boston for the 2025 AI-Driven Drug Discovery Summit, the leading global event focused solely on AI in drug discovery and development.  November 18–20 - Thomas M. Menino Convention & Exhibition Center With the pace of AI moving from research to real-world pipelines, this year’s summit hits at the right time. Expect deep dives into foundation models, multimodal biology, and how AI-native platforms are changing what “drug development” even means. Salt’s building the infrastructure to make these models actually work in production across target ID, preclinical design, and trial acceleration. If you’re heading there, we’d love to connect. Contacts: Toby Sayre, Dan Sheikh See you in Boston. Learn more: https://lnkd.in/gid5t8vr #AIDrugDiscovery #AIInBiotech #SaltAI #TargetDiscovery #AIMedicin #AIDDD