Director of AI ML

📋 Description

• • Set the strategic north star for Bayesian Health’s AI/ML organization, translating our life-saving mission into a concrete technical roadmap that spans real-time clinical decision support, predictive risk modeling, and adaptive learning systems deployed in hundreds of hospitals nationwide.
• • Build, mentor, and scale a multidisciplinary team of data scientists, ML engineers, and research scientists from the ground up—establishing hiring rubrics, leveling frameworks, and a culture of psychological safety where the best ideas win and every model is judged by the lives it saves.
• • Architect and own the end-to-end ML platform: streaming ingestion of HL7/FHIR feeds, feature stores that respect PHI boundaries, experiment-tracking infrastructure, real-time inference services with sub-second latency, and automated model monitoring that catches drift before a single patient outcome is affected.
• • Roll up your sleeves and stay technical—prototype novel deep-learning architectures for time-series forecasting on sparse, irregular clinical data; debug race conditions in our Kafka-to-TensorFlow-Serving pipeline at 3 a.m.; and run counterfactual simulations to quantify how many sepsis deaths our latest model would have prevented last quarter.
• • Partner elbow-to-elbow with physicians, nurses, and quality-improvement leaders to translate messy clinical workflows into crisp ML problem statements, then translate model predictions back into intuitive, actionable UI cues that fit seamlessly into Epic and Cerner workflows.
• • Establish rigorous evaluation protocols that go beyond AUROC—design prospective clinical trials, work with our biostatisticians on propensity-weighted analyses, and publish results in journals like Nature Medicine that demonstrate causal impact on mortality and length-of-stay.
• • Own the P&L of the AI/ML org—forecast cloud spend vs. lives-saved ROI, negotiate enterprise GPU contracts, and make build-vs-buy decisions on everything from feature stores to AutoML tooling while staying scrappy and resource-constrained.
• • Evangelize our work externally—present at NeurIPS, HIMSS, and AMIA; advise hospital CIOs on AI governance; and serve as the public face of Bayesian Health’s technical brand, turning complex ML concepts into stories that inspire clinicians and investors alike.
• • Champion responsible AI practices—bias audits on under-represented populations, fairness constraints in model training, and transparent documentation that satisfies FDA SaMD, HIPAA, and hospital IRB requirements.
• • Foster a feedback-rich environment where post-mortems are blameless, every failed experiment is a learning opportunity, and the team celebrates incremental wins like shaving 50 ms off inference latency because those milliseconds matter when a patient is crashing.
• • Navigate the chaos of an early-stage startup—reprioritize weekly as new hospital partnerships come online, balance technical debt against speed-to-impact, and maintain zen-like calm when the ETL pipeline breaks the night before a big customer demo.
• • Leave a legacy—create playbooks, open-source tools, and architectural patterns that future ML leaders at Bayesian Health will reference for years, knowing that your fingerprints helped turn reactive sick-care into proactive, data-driven healthcare.