AI Officer - Scientific AI Harnesses, Agents & Tooling (ref. AD/26/11)
11 hours ago
Barcelona
IRB Barcelona is launching a new Artificial Intelligence group and is looking for AI Officer to help build it. The group’s mission is not to train frontier models, but to make them genuinely useful for science: turning powerful AI models and the institute’s data into reliable, reusable systems that accelerate biomedical discovery. The successful candidate will design, build and maintain standardised AI ecosystems — what we call a “harness” — around large language models and other AI models: the interfaces, knowledge and retrieval layers, tools and connectors, reusable workflows, evaluation and safeguards that let research groups apply frontier AI in a rigorous, reproducible way. Positions are embedded with research groups and core facilities, so the work has direct, visible impact on ongoing science. We are looking for people with a solid software and engineering foundation and a deep interest in AI architecture. The emphasis of the role is on building and integrating systems; a smaller part involves selecting, adapting, fine-tuning and evaluating existing open or commercial models — not developing new model architectures. Previous experience deploying AI in production is an advantage, but it is not essential. • Design, build and maintain the “harness” around AI models: agents, retrieval-augmented generation (RAG), tools, APIs, connectors and interoperability protocols (e.g. MCP), reusable skills and workflows., • Connect research groups’ data, databases and literature to models through RAG and knowledge/graph layers, so answers are grounded in verifiable, institutional evidence., • Select, adapt, fine-tune and evaluate existing models (open-source and commercial), rather than training new architectures from scratch., • Build evaluation harnesses — benchmarks, metrics, human-in-the-loop review and guardrails — so nothing is deployed without measurement and oversight., • Deploy and operate AI models, agent environments and tool integrations across on-premises infrastructure and commercial AI platforms; monitor quality, availability, latency, usage and cost; and ensure that components are versioned, observable and recoverable., • Ensure reproducibility, data governance and FAIR principles: provenance, versioning, traceability, access control and responsible-AI practices, especially with sensitive or clinically relevant data., • The team works in close collaboration with research groups and core facilities, ensuring that the systems developed address real scientific use cases and have a direct, visible impact on ongoing research., • Contribute to institute-wide AI pilots (for example, literature-monitoring agents, analytical toolkits and coding assistants) and to the responsible roll-out of AI tools., • Produce technical documentation, examples, tutorials and internal training., • Deploy, configure and maintain AI-assisted scientific environments based on tools such as OpenCode, Cline, Open WebUI and others., • Create reusable skills that encode scientific procedures, analytical workflows, validation criteria and domain-specific instructions., • Develop MCP-compatible tools and connectors that expose scientific software, databases, internal services and analytical methods to AI agents through controlled interfaces., • Maintain registries of models, tools, skills, prompts and workflows, including their versions, permissions, intended uses and evaluation status., • Design model-routing and provider-abstraction mechanisms that allow workflows to use locally deployed models or commercial services from Anthropic, OpenAI and Google according to scientific quality, data sensitivity, latency, availability and cost., • Integrate commercial model APIs, including tool calling, structured outputs, multimodal inputs, usage controls and provider-specific capabilities, while avoiding unnecessary dependence on a single vendor., • Collaborate with the Scientific Computing, IT, data-governance and research teams, and evaluate new technologies that improve usability, performance or reproducibility. We do not expect candidates to already have extensive experience building AI systems in production. The essential requirements are: • A background in mathematics, physics, computer science, software/computer engineering, data science or a related quantitative discipline., • Basic knowledge of Linux and command-line environments., • Working knowledge of English., • Genuine interest in AI architecture: agents, retrieval-augmented generation, tool/model integration and how AI systems are made reliable., • Ability to work as part of a technical team and to communicate with researchers., • A general understanding of machine learning and modern AI, including large language models., • Ability to troubleshoot technical problems in a structured way and to build things that other people can use., • Programming skills in at least one language, preferably Python (others such as JavaScript/TypeScript, Go, C++ or similar are welcome). Experience or knowledge in any of the following would be valuable, but is not mandatory: • Experience designing AI harnesses that combine models, tools, skills, workflows and human supervision., • OpenCode, Cline, Open WebUI or comparable AI-assisted environments., • Model Context Protocol (MCP) and development of tools or connectors for AI agents., • Integration with Anthropic, OpenAI and Google APIs., • Deployment and use of open-weight or locally hosted language and multimodal models., • Model routing between on-premises and commercial services based on quality, privacy, latency and cost., • Retrieval-augmented generation, semantic search, vector databases or knowledge graphs., • Evaluation of agents and AI workflows, including benchmarks, regression tests and human review., • Detection of hallucinations, incorrect tool use, prompt injection and data leakage., • Git, automated testing, APIs, containers and collaborative software-development practices., • Scientific or biomedical applications such as bioinformatics, genomics, drug discovery or imaging., • FAIR data, provenance, traceability and reproducible research. We are looking for candidates with strong technical curiosity and a systems mindset — people who enjoy understanding how AI models, data and infrastructure fit together, and who take pride in building tools that others rely on. We value rigour and responsible, reproducible AI, good communication with scientists, and a willingness to learn. Candidates do not need to meet every item in the “you might also have” section to apply. To be a part a part of something meaningful: We are a biomedical research centre that tirelessly works to achieve a society free of diseases. Our mission is to carry out research of excellence to cure cancer and other diseases linked to ageing. Good Working conditions: Permanent Contract (Contrato de actividades científico-técnicas). Employed in compliance with Spanish legislation and regulations under a full-time contract. Employees receive the benefits of the Spanish Social Security system covering sickness, maternity/paternity leave and injuries at work. Competitive salary: commensurate with experience and qualifications. International Environment: we are home to nearly 500 researchers, technical and support staff from 45 countries. Multidisciplinary Research: Research of excellence at the unique interface between biology, chemistry, and structural and computational biology. Continuous Training: Transversal and general training is offered to our community (examples include research seminars, conferences, training in time management, scientific integrity, gender equality, health and safety, languages courses and emotional wellbeing). Relocation & Visa Assistance: Trained staff provides support on visa paperwork, administrative requirements, and practical information on living and working in Barcelona, etc. Social Life: We like to have fun. We promote several activities, like a Football League, an annual skiing trip, a Mountaineering Club, a Running Club and "Cool-off sessions", where members meet informally for drinks, snacks and scientific discussions. Our recruitment is open, transparent and merit-based, and all applicants compete on the same terms. We are an Equal Opportunity Employer and all qualified applicants are considered for employment without regard to race, colour, religion, age, sex, sexual orientation, gender identity, nationality, ethnic origin or disability. • Deadline for applications: 19/08/2026 (If no suitable candidate is found, the deadline will be extended), • Number of positions available: 1, • Selection process:, • Pre-selection: Will be based on CV, motivation letter & experience,, • Interviews: Short-listed candidates will be interviewed., • Job offer: Will be sent to the successful candidate after the interview. IRB BARCELONA endorses the Requirements and Principles of the European Charter for Researchers, the Code of Conduct for the Recruitment of Researchers, and Open, Transparent, Merit-based recruitment promoted by the European Commission and follows Equal Opportunities policies. On 9 December, 2014, IRB Barcelona was awarded the "HR Excellence in Research" logo. This recognition reflects the commitment of the Institute to the continuous improvement of its human resources policies in line with the Charter & Code. The Institute works to ensure fair and transparent recruitment and appraisal procedures. Recruitment Policy. Note: The strengths and weaknesses of the applications will be provided upon request. Created in 2005 by the Generalitat de Catalunya (Government of Catalonia) and the University of Barcelona, IRB Barcelona, a Severo Ochoa Centre of Excellence since the first call in 2011, has received the accreditation four consecutive times. The institute is devoted to conducting research of excellence in biomedicine and to transferring results to clinical practice, thus improving people’s quality of life, while simultaneously promoting the training of outstanding researchers, technology transfer, and public communication of science. Its 27 laboratories and 9 core facilities address basic questions in biology and are oriented towards diseases such as cancer, metastasis, Alzheimer’s, diabetes, and rare conditions. IRB Barcelona is an international centre that hosts 500 members from 45 nationalities. It is located in the Barcelona Science Park. IRB Barcelona forms part of the Barcelona Institute of Science and Technology (BIST) and the “Xarxa de Centres de Recerca de Catalunya” (CERCA). #J-18808-Ljbffr