12 months ago
Base Salary
$190k - $260k/yr
Responsibilities
- Own end-to-end applied research from problem framing and experiment design through production deployment and real-world impact monitoring
- Set technical direction across LLMs, retrieval, and multimodal modeling and conduct ablations and error analyses that influence product decisions
- Build evaluations connecting offline metrics to online outcomes and define thresholds, monitoring, and rollback procedures
- Partner with engineering, product, clinicians, and domain experts to align data, success criteria, and timelines
- Mentor peers and establish standards for reliability, safety, and documentation
- Improve data, training, serving, and observability platforms to accelerate iteration and reproducibility
- Explore strategic computer vision and vision-language research directions
Requirements
- MS or PhD, or equivalent research experience, in Computer Science, Electrical Engineering, Computational Linguistics, Biomedical Informatics, or a related quantitative field
- At least 7 years of applied ML research experience, or a PhD plus 5 years, or equivalent evidence of Staff-level impact
- Depth in one or more of LLMs and NLP, computer vision, speech, recommendation/ranking, retrieval, or multimodal modeling
- Strong experimental rigor, including hypothesis framing, offline-to-online linkage, calibration, stratified analyses, and decision-influencing ablations
- Proven ability to take models into production
- Hands-on experience with PyTorch and modern experiment or operations tools such as MLflow, Databricks, Ray, or similar
- Ability to design for observability and rollback and clearly document technical decisions
- Strong written and verbal communication skills, including design documentation and explaining complex ideas to non-specialists
- Experience with health data such as EHR or imaging is preferred
- Experience with clinical NLP or LLMs, computer vision, speech, retrieval, or multimodal modeling is preferred
- Experience shipping and measuring monitored production models, with external or multi-site validation a plus
- Experience integrating workflows with EHR, RIS, PACS, or reporting systems is preferred; PowerScribe or Dragon exposure is helpful
- Experience with safety and governance in sensitive domains, including PHI handling and HIPAA- or FDA-adjacent environments, is preferred
- Technical mentorship, research-culture contributions, publications, or impactful open-source work are preferred
- Experience with PyTorch and ML operations tools such as MLflow, Databricks, Ray, or Triton is preferred
Benefits
- Comprehensive medical, dental, vision, and life insurance
- HSA with employer match, FSA, and DCFSA
- 401(k)
- 11 paid company holidays
- Flexible PTO policy
- Annual company-wide offsite and periodic team offsites
- Annual equipment stipend
- Applicants may work across the United States, with a preference for the role to be based in the San Francisco office
Tech Stack
Categories
AI ResearchML Engineering
About Rad AI
Rad AI builds AI software for radiologists and health systems to streamline reporting and radiology workflows, integrating with PACS and EHR systems. Products include generative AI tools for report drafting and quality improvement used by radiology groups and hospitals such as Cone Health, Geisinger, and Henry Ford Health. Founded in 2018, the privately held company is headquartered in San Francisco.