# Sangfroid Labs > Proving ground for tools designed to operate during crisis and emergency. Sangfroid Labs studies complex systems in and near emergencies, and builds tools designed to operate during crisis. Founded by Daniel Dworkis, MD PhD. Contact: sangfroid@agentmail.to Website: https://sangfroidlabs.com ## Pages - [Home](https://sangfroidlabs.com): Landing page - [About](https://sangfroidlabs.com/about.html): About Sangfroid Labs - [Projects](https://sangfroidlabs.com/projects.html): All projects - [Emergency Mind](https://www.emergencymind.com): Emergency Mind — cognitive performance and emergency medicine ## Projects ### HALOSIMR (HaloSim for R) Year: 2026 Tags: software, R, training, HALO, simulation, halosim Link: https://github.com/EmergencyMind/halosimr An R package to simulate HALO (High Acuity Low Occurrence) event exposure across a provider population, model how readiness decays between exposures, and estimate how training programs change it. ### Emergency Alignment for Human-AI Teams Year: 2026 Tags: AI safety, emergency medicine, alignment Link: https://www.psychologytoday.com/us/blog/the-emergency-mind/202608/emergency-alignment In a crisis, our values temporarily change. Emergency alignment asks whether human-AI teams can make that transition together, and what happens when they can't. ### Cross-Model Variability in Large Language Model Triage Behavior for Potential Stroke Symptoms Year: 2026 Tags: AI safety, emergency medicine, triage, public health Preprint: https://www.medrxiv.org/content/10.64898/2026.05.22.26353904v1 Frontier LLMs under-responded to potential stroke like symptoms, and showed substantial model-to-model differences in response to changing symptom descriptions. As individuals increasingly turn to LLMs for medical advice, which model they use and what words they choose could lead to drastically different informal triage guidance. ### Building in Calm, Deploying in Crisis Year: 2026 Tags: emergency medicine, design, AI safety Link: https://www.linkedin.com/pulse/building-calm-deploying-crisis-dan-dworkis-md-phd-p9afc/ Most AI systems are built in calm but must perform in crisis. Three structural gaps explain why current AI falls short in emergency medicine: training on theory not practice, output designed for calm cognition not crisis, and systems built for individuals not teams. ### HALOSIM (Web App) Year: 2026 Tags: software, training, HALO, halosim App: https://sfl-halosim.streamlit.app/ Software for HALO (High Acuity Low Occurrence) event exposure and training simulation for emergency teams. ### Prompt-Injected State Descriptions Shift Surgical Airway Decision Timing in a Large Language Model Year: 2026 Tags: simulation, AI safety, emergency medicine, decision making Preprint: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6409858 Modeling LLM decision-making in CICO (Can't Intubate, Can't Oxygenate) emergencies. Prompt-injections focused on altered affective and cognitive states substantially changed LLM decision making. ### Silicon Teammates: How Human-AI Teams Make Hard Decisions Year: 2026 Tags: emergency medicine, AI safety, decision making Link: https://www.psychologytoday.com/us/blog/the-emergency-mind/202603/silicon-teammates-how-human-ai-teams-make-hard-decisions A human-AI team is a dyad: two agents and the relationship between them. High-performing partnerships are built in training before crisis, not improvised during one. ### The Emergency Mind Book Year: 2021 Tags: book, cognitive performance, emergency medicine Link: https://www.amazon.com/dp/B09447QCYG Wiring Your Brain for Performance Under Pressure. How to train mentally to perform at your best when you're needed the most. ### Emergency Mind Podcast Year: 2020 Tags: podcast, cognitive performance, emergency medicine Link: https://podcasts.apple.com/us/podcast/the-emergency-mind-podcast/id1480163086 Conversations on performance, decision-making, and the human side of emergency care.