Medical AI Director at Centaur (W19) $4K - $7K / monthly Quality-first data labeling for model evaluation and training Remote Full-time US citizenship/visa not required 3+ years About Centaur At Centaur, we label data at scale to enable breakthroughs in medical AI. Our novel approach uses collective intelligence to aggregate medical opinions from experts and students all over the world. Our labeling customers include leading AI companies and prominent research organizations. We're funded by SignalFire, Matrix, Accel, and Y Combinator. About the role Centaur Overview: Centaur AI is the modern data labeling solution driving AI innovation across the healthcare industry. From agile startups and cutting-edge research labs to global pharmaceutical giants and Big Tech, we empower the entire spectrum of AI. Our expert crowd of 50k+ labelers, combined with our gamified DiagnosUs app for motivation and quality control, ensures scalable, high-quality labels are delivered in as little as 24 hours. We are backed by leading investors including Y Combinator, Matrix, and Accel, and recently closed a Series B led by SignalFire. **Full-Time Position Overview: **You’ll be the clinical lead and quality owner for complex medical labeling across imaging and text, with a radiology-first focus. You will design pilots and gold standards, build review/QC pipelines, and translate clinical nuance into clear, scalable labeling rules. You’ll also scope and manage offshore radiologist teams/partners, and act as a medical point-of-contact on select client/physician conversations. Internally, you’ll author training materials and collaborate closely with supervisors/Project Managers to kick-off projects faster, define performance management metrics, and maintain accuracy at scale. The ideal candidate is: A non-US radiologist (LATAM/Europe for maximum EST overlap) with 5+ years hands-on experience on annotation/QA who writes crisp guidelines, reasons well on edge cases, and communicates clearly with both technical and GTM teams. Main Responsibilities Gold standards & pilots: Run time-trials, set acceptance criteria, define edge-case policies, and establish baseline throughput/quality targets. Clinical point-of-contact: Join select client/physician calls; convert clinical nuance into crisp, scalable labeling rules. Training guidelines (“Centaurized” SOPs): Co-author step-by-step instructions with customers, incl. glossary/taxonomy, positive/negative examples, edge cases, and arbitration paths. Reviewer/QC pipelines: Design arbitration workflows, gold-standard sampling & update loops, error taxonomy, escalation ladders, retraining triggers, and safety checks. Quality monitoring: Define QC metrics, set performance-drift alerts, analyze trends, and drive corrective actions. Training & coaching: Upskill supervisors/annotators on tricky edge cases; turn lessons learned into reusable playbooks. Radiology team leadership: Scope, onboard, and manage offshore radiologists/partners; align capacity, quality, and cost; maintain a shared intake of capabilities for rapid staffing. Decision-quality evaluation: Audit outputs for clinical safety, accuracy, and reasoning; convert error patterns into rules and compact decision trees/heuristics. Qualifications Clinical training: Radiologist (MD) trained/practicing outside the US strongly preferred (LATAM/EU); other clinicians considered only with strong annotationQA track record. Education: Medical degree required with certification; open to residents, practicing clinicians, or those with extensive clinical experience. Timezone: Able to overlap ~6 hrs with US Eastern Time (EST); Other timezones will be reviewed on a case-by-case basis. Annotation/QA Experience: 5+ years experience with imaging guidelines, gold standards, sampling, arbitration, and QA metrics. Team leadership: Experience coordinating radiologists or clinical vendors (intake, scoping, coaching, QA). Tooling: Comfortable with annotation platforms, spreadsheets/BI, and lightweight automations. Domain exposure (pluses): CT/MRI/Ultrasound/IVUS; chest/thoracic & oncology; medical coding knowledge; other medical fields a plus. Core Competencies Ability to make, and explain, consistent labeling judgments on edge cases. Ability to author clear, unambiguous guidelines/SOPs. Experience designing QA and metrics, incl. pilots/time-trials, sampling, and KPI tracking for accuracy & throughput. Ability to communicate in fluent English with PMs and customers on clinical/annotation topics. Experience evaluating clinical reasoning & safety and translating recurring errors into decision trees/heuristics. (STRONG PLUS) Don’t meet every requirement? At Centaur AI, we are dedicated to building a diverse, inclusive, and authentic workplace. So, if you’re excited about this role but your experience doesn’t perfectly align with every qualification in the job description, we encourage you to apply anyway. You might just be the right candidate for this or other roles! Why join Centaur? Impact: Shape how state-of-the-art medical AI is built—your rules and QA systems become the standard. Growth: Lead a radiology-first program, manage offshore teams, and codify playbooks used across customers. Flexibility: Remote role with meaningful EST overlap; this role will start as a contractor but with a path towards full-time employment available. Compensation: Competitive salary. Modality: This is a full-time position (40 hrs/week) Fabulous company culture centered around our values: Act with integrity: We always do what is right for ourselves, our team, our customers, our community and the world. Create change: Believe in ideas that seem unrealistic or unconventional. Take initiative to improve the comptaoany and the world, unprompted. Keep learning: Seek out opportunities to ask questions, test assumptions, learn, and improve. Stay curious. Every voice counts: Collective intelligence is core to our company. Speak your mind. We are more than the sum of our parts. Our Commitment to Diversity Centaur AI is proud to be an Equal Opportunity Employer. We are deeply committed to building and supporting a diverse team through inclusive, equitable hiring practices. Technology We use React.js and Swift on the frontend, and Python and Swift (Vapor) on the backend. Our data store is PostgreSQL, and we're hosted on AWS. Centaur faces a wide variety of difficult technical problems, from efficiently serving millions of images and videos to our users, to ensuring our platform is secure and our valuable data sets remain private, to performing complex analysis on our users' performance and building machine learning models to take our decision making capabilities to the next level.