About Gray SwanGray Swan protects organizations from emerging AI security threats. We build real-time threat detection, automated validation, and adaptive defenses for AI labs and enterprises. We're a team of ~25 people, well-funded, and growing fast.The RoleAs a Machine Learning Researcher at Gray Swan AI, you will work at the boundary between research and production — developing new approaches to adversarial testing, model evaluation, and robust inference that directly inform how secure AI systems are built and deployed.AI security is not a solved problem. This role is fundamentally about research: designing experiments, prototyping novel methods, analyzing results empirically, and translating findings into real systems that withstand adversarial pressure. You will collaborate closely with engineering and platform teams to see your ideas through from prototype to production impact.What You'll DoDesign and develop novel ML approaches to adversarial testing, model evaluation, and robust inference.Build and deploy ML models that meet real-world performance and scalability requirements.Design experiments, analyze results empirically, and communicate findings through publications and internal research reports.Develop and advance methodologies for controlling, monitoring, and analyzing ML models in production environments.Translate research ideas into scalable AI systems deployed in real-world, adversarial settings.Work closely with cross-functional teams to ensure research outcomes inform production systems.Who You AreEducationBachelor's degree in Computer Science, Machine Learning, Engineering, or a related technical field.A Master's or PhD in a relevant technical field is strongly preferred, especially with a focus on machine learning and AI safety.ExperienceExperience building and deploying ML models and systems.Demonstrated expertise in designing, training, and deploying deep learning models, particularly with PyTorch.Strong Python programming skills; C++ preferred.Experience developing scalable ML pipelines and integrating with cloud infrastructure (AWS, GCP, or Azure).ML research experience: building research prototype systems, designing experiments, empirical analysis of results, and communicating results via publications.Core Technical SkillsNeural network architectures, including transformers, sequence models, and other state-of-the-art approaches.Strong algorithmic problem-solving skills and knowledge of ML theory and optimization.Data preprocessing, transformation, and handling large-scale, multi-modal datasets.Bonus Points If You HaveExperience with LLMs (training, fine-tuning, or analyzing), synthetic data generation, and AI safety or security work.AI safety practices: model validation, robustness testing, and continuous monitoring for safety and security incidents.Familiarity with AI safety and security assessments and adversarial testing.You'll Thrive Here IfYou are genuinely excited by the intersection of research and engineering, and want to both develop new AI safety ideas and see them running in real systems.You are motivated by real-world impact and want your work to directly influence how major AI companies deploy models right now (we work with many of the leading AI labs).You are eager to deepen your AI safety expertise by working alongside a team that includes some of the most respected and influential thinkers in the field.You thrive in a fast-paced, dynamic startup environment where ambiguity is expected.You bring strong collaboration and problem-solving skills, with a focus on driving meaningful, lasting impact.Compensation & BenefitsWe offer a competitive compensation package designed to reward impact and incentivize growth. Our compensation philosophy is informed by our current valuation and recent industry data.Salary: $183,000-$278,000Equity: Competitive equity packageBenefits:401k with up to 4% matching28 days annual leave (vacation + holidays)Health, dental, and vision coverageCatered lunches (Pittsburgh office)Flexible work arrangementsVisa sponsorship available for exceptional candidates
