LogoFresh Jobs Ahead
Johnson & Johnson
Johnson & Johnson
H1B vizasi homiysiFarmatsevtika mahsulotlariSog‘liqni saqlashTibbiy sug‘urta
Johnson & Johnson innovatsion dori vositalari va tibbiy texnologiyalar yo‘nalishlarida faoliyat yurituvchi global sog‘liqni saqlash kompaniyasidir. Uning jamoalari onkologiya, immunologiya, nevrologiya, maxsus oftalmologiya, intervensional tibbiyot va ortopediya sohalarida yechimlar ishlab chiqib, murakkab tibbiy ehtiyojlarni hal etish uchun fan va texnologiyani birlashtiradi. Kompaniya butun dunyo bo‘ylab tibbiyot xodimlari, bemorlar va jamoalarga xizmat ko‘rsatish bilan birga, sog‘liqni saqlashda tenglik, ekologik barqarorlik, xilma-xillik va inklyuzivlikka ham katta e’tibor qaratadi.
LinkedIn
Kompaniyadagi ish o‘rinlari

Senior AI Safety Scientist – Johnson & Johnson, Hybrid Madrid

Lead AI safety research for Johnson & Johnson’s pharmaceutical R&D platform, focusing on provable controls, adversarial assurance, and safety-centered architectures. Hybrid position based in Madrid or Barcelona.

ESIspaniya
55 400 € – 87 860 € / yil
GibridTo‘liq stavkaSeniorTalab Qilinadigan Til: inglizchaMashinaviy o‘qitish muhandisiSunʼiy intellekt tadqiqot olimi

Tavsif

  • Establish deterministic, explainable controls across agentic workflows.
  • Create machine-readable policies for agent behavior, tool use, data access, and information flow.
  • Deploy auditable enforcement for high-impact agent actions, including human-approval routes.
  • Provide accessible policy interfaces for owners to create, test, and maintain controls.
  • Develop continuous red-team approaches combining AI threat models with pharmaceutical R&D failure scenarios.
  • Integrate adversarial testing into the GenAI Platform across models, agents, tools, and complete workflows.
  • Set evaluation methods, adjudication rules, and evidence thresholds for safety properties.
  • Research and prototype safety-centered architectures built on layered technical controls.
  • Assess pre-training data interventions and post-training techniques such as supervised fine-tuning, preference optimization, and safety tuning.
  • Convert scientific, quality, privacy, and regulatory obligations into testable system specifications.
  • Establish accountability measures and evidence requirements for AI safety claims.
  • Maintain traceable, reconstructable control decisions for audit evidence.
  • Collaborate with R&D scientists, platform engineering, security, privacy, legal, and quality teams.
  • Guide the scientific work of a small cross-functional team and represent it through publications and standards contributions.
  • Develop prototypes, define technical direction, and transition results into the platform with engineering partners.

Talablar

  • PhD in computer science, AI/ML, applied mathematics, or a closely related technical discipline is required.
  • At least one year of post-PhD research or industry experience developing and deploying AI/ML, autonomous-agent, or security-critical systems.
  • Advanced practical knowledge of agentic AI, including foundation models, retrieval-augmented generation, tool orchestration, memory, planning, multi-agent frameworks, and related failure modes.
  • Research or applied experience in AI safety and alignment, including supervised fine-tuning, preference optimization, safety tuning, adversarial evaluation, interpretability, or scalable oversight.
  • Expertise in at least one area: policy-as-code and authorization; adversarial machine learning and AI red teaming; information-flow or data-access control; provenance and lineage; or formal specification and verification.
  • Strong AI engineering ability, including prototyping evaluation and control techniques, running reproducible experiments, and producing production-ready code.
  • Excellent written and spoken communication skills.
  • Scientific rigor in describing model and agent behavior, uncertainty, limitations, failure conditions, and supporting evidence.
  • Published or otherwise recognized work in relevant leading journals or conferences is preferred.
  • Experience adapting foundation models for scientific domains and evaluating their safety properties is preferred.
  • Practical experience with policy engines, authorization languages, or formal policy specifications is preferred.
  • Familiarity with MITRE ATLAS, OWASP LLM guidance, NIST AI RMF, ISO/IEC 42001, or the EU AI Act is preferred.
  • Experience with data classification, de-identification, and privacy-preserving machine learning is preferred.
  • Experience deploying AI in life sciences or another regulated, high-stakes environment is preferred.
  • Familiarity with GxP, data integrity, and validation of AI-enabled computerized systems is preferred.
  • Experience assessing third-party AI safety, evaluation, or control technologies is preferred.
  • Experience creating reusable AI safety capabilities adopted by multiple teams is preferred.

Imtiyozlar

  • Annual bonus with a target determined by pay grade and location, based on individual and company performance.
  • Paid vacation days.
  • At least 12 weeks of parental leave.
  • Bereavement leave.
  • Caregiver leave.
  • Volunteer leave.
  • Well-being reimbursement.
  • Financial, physical, and mental health programs.
  • Service anniversary and recognition awards.
  • Insurance coverage for employees and, in some locations, eligible dependents.

O‘xshash ish o‘rinlari