Agentic AI Scientist | Informatics Engineering (or Related Area) | Intelligent Systems Team (Job Ref.: AICOS_Jobs_2025_27) | Porto, Portugal | English Version

Job Title: Agentic AI Scientist Job Ref: AICOS_Jobs_2025\_27 General scientific field: Engineering Specific scientific field: Informatics Engineering (or related area) • Associação Fraunhofer Portugal Research opens an international call for PhD holder researchers, under a non-fixed work contract with an expected duration of 12 months. This Announcement Is For One Researcher Position To Conduct Scientific Research Activities Within The Following Projects And Project Proposals (if Approved) • AI4CarOps – AI-Powered Vehicle Reception Operations for Car Workshops, with Notice No. MPr-2023-07 SIID I&D Empresarial and Project Reference No. 18234 - COMPETE 2030 Copromoção. • FibFlow – Financial Intelligent Business Flow; with Notice No. MPr-2023-08 SIID I&D Empresarial and Project Reference No. 20856 - COMPETE 2030 Copromoção. • Microelectrónica – Agenda Microeletrónica, with Notice No. 01/C05-i01/2021 and Project Reference No. 19 - NextGenerationEU (PRR) Copromoção. • NextGenAI – Center for Responsible AI, with Notice No. 01/C05-i01/2021 and Project Reference No. 62 - NextGenerationEU (PRR) Copromoção. • DigitalFOOD – Fostering AI adoption in the Portuguese Food Industry, Project Proposal submitted within Notice No. COMPETE2030-2025-4 - COMPETE 2030 SIAC Digitalização. • RAICC – Responsible AI Call Center, Project Proposal submitted within Notice No. MPr-2025-01 SIID I&D&I Empresarial - COMPETE 2030 Copromoção. The research will focus on the scientific area of Large-Language Models, Retrieval-Augmented Generation (RAG) and Agentic AI, with the goal of developing reliable, policy-compliant AI methodologies for critical domains, automating complex data-intensive tasks, and establishing robust evaluation protocols for AI response quality. The Work Plan Includes The Following Tasks • Research, implement, and validate pioneering Retrieval-Augmented Generation (RAG) methodologies in real-world settings. This includes creating policy-sensitive RAGs that comply with regulatory documents and integrating semi-automated compliance checks. • Automate repetitive business tasks by extracting and structuring data from documents and designing AI models to provide categorization recommendations based on the extracted information. • Continuously assess complex documentation using advanced Large-Language Models (LLMs) to provide context-aware risk assessments, generating dynamic reports and risk scores. • Develop specific metrics and protocols to evaluate the quality of AI-based systems. • Conduct AI training and the creation of proof-of-concepts (PoCs) for new applications. • Applicable Legislation: Decree-Law no. 57/2016 of August 29th, amended by Law 57/2017 and Regulatory Decree No. 11- A / 2017 which approved the doctorate hiring regime destined to stimulate scientific and technological employment for all knowledge areas (Regime Jurídico do Emprego Científico - RJEC); Portuguese Labor Code, approved by Law 7/2009 of February 12, in its actual form. • Pursuant to article 13 of RJEC, the tender selection panel is formed by: • Inês Nunes de Sousa Soares (PhD; Head of Department at Associação Fraunhofer Portugal Research; Chairman of the Jury). • Luís Filipe Caeiro Margalho Guerra Rosado (PhD; Permanent Member). • André Valério Raposo Carreiro (PhD; Permanent Member). • The workplace shall be at Associação Fraunhofer Portugal Research – Rua Alfredo Allen 455/461, 4200-135 Porto, Portugal – which may involve travel in Portugal or abroad, in the service or representation of the employer. • Monthly remuneration to be paid is that set by subheading a) nr.1 article 15 of RJEC and article nr 2 of the Regulatory Decree nr. 11-A/2017, corresponding to level 38 of the Tabela Remuneratória Única, approved by Order no. 1553-C/2008 of December 31st, i.e. 2.622,59 Euros, corresponding to index FhP_SSct_2 (Senior Scientist) of Associação Fraunhofer Portugal Research 's Careers Policy. • Application can be submitted by any national, foreign and stateless candidate(s) holding a doctorate degree in Informatics Engineering or related area, and a scientific and professional background that aligns with the specific activities described below (see item 7). In case the doctorate degree was awarded by a foreign higher education institution, it must comply with the provisions of Decree-Law no. 66/2018 of 16 October, and all formalities established therein must be complied with at the signature of work contract. • The tender admission general requirements are those defined in the previous point (see item 6), and the additional specific requirements: • Experience in Generative AI, Large Language Models, reinforcement learning, recommender systems, and multi-agent systems. • Research experience in Responsible AI. • Experience with designing and architecting end-to-end evaluation pipelines and benchmarking for AI-based systems. • Relevant hands-on experience in AI solutions design, development and implementation. • Teaching, supervision, and mentoring experience relevant to the scientific domain. • Successful experience in research proposal writing. • Programming proficiency in Python and experience with machine learning frameworks such as PyTorch and TensorFlow. • Proficiency in English and Portuguese. • Pursuant to article 5 of RJEC, the selection is to be made based on the evaluation of the scientific and curricular career of the candidate. • Scientific and curricular career evaluation focuses on relevance, quality and up-to-datedness: • Of scientific, technological, cultural or artistic production in the last five years, considered most relevant by the candidate. • Of research activities, fundamental or applied science, developed in the last five years, considered of higher impact by the candidate. • Of knowledge extension and dissemination activities developed in the last five years specifically related to the promotion of culture and scientific practices, as deemed most relevant by the candidate. • Of activities of science management, technology and innovation programs, as well as experience in observing and monitoring the scientific and technological system or higher education, in Portugal or abroad. • The five-year period mentioned above can be extended by the panel, if requested by the candidate, whenever the suspension of scientific activities is reasoned by socially protected grounds like paternity leave, long-term serious illness, and other legal situations of unavailability to work. • Evaluation criteria are the following: The evaluation of the Curriculum Vitae of the candidates, in particular the scientific merit and research experience will take into consideration the elements presented for the last five years of activity falling within the scope of the specific areas of the tender (see item 1) and will be performed according to the following criteria: 11.1 Integrated assessment of the curriculum trajectory of the candidate, based on an overview of their scientific merits, namely: • Overall scientific coherence of the Curriculum Vitae, including participating in research projects and supervision – 15%. • Diversity and quality of scientific indicators, including relevant published peer-reviewed articles in the specific areas of the announcement (see item 1) – 5%. 11.2 Relevant Experience, Proved In Curriculum Vitae, In • Generative AI, Large Language Models, reinforcement learning, recommender systems, and multi-agent systems – 10%. • Research in Responsible AI – 10%. • Designing and architecting end-to-end evaluation pipelines and benchmarking for AI-based systems – 10%. • Hands-on AI solutions design, development and implementation – 10%. • Teaching, supervision, and mentoring experience relevant to the scientific domain – 10%. • Research proposal writing with successful outcomes – 10%. • Python and machine learning frameworks such as PyTorch and TensorFlow – 10%. • Communication with project partners in English and Portuguese – 10%. • In the case of admitted candidates whose evaluation does not differ more than 10% from that obtained by the best positioned candidate, the jury will interview these candidates. This will be aimed at obtaining clarifications about the curricular elements and additional information, as well as to evaluate the attitude, profile and motivation of the candidate. The final score of each jury member is obtained by the following assessment: 90% scientific and curricular career evaluation and 10% interview. • The candidates’ final classification system shall be given based on a scale 0 to 100. • The panel shall deliberate by means of roll-call vote justified under adopted and disclosed selection criteria, with no abstentions allowed. • Minutes of panel meetings shall be executed and include a summary of all occurrences of said meeting, as well as of all votes casted by the members and respective reasoning. • After selection criteria application, the panel shall prepare a sorted list of approved candidates and respective classification. • The panel’s final decision shall be validated by the leader of the institution, who is also in charge of deciding about the hiring. • Application formalization: 18.1 The applications are formalized at the electronic address jobs@fraunhofer.pt following documents in a digital form, in PDF format: • Curriculum Vitae. • Motivation Letter. • Qualifications Certificate. 18.2 The application period is from 16-10-2025 to 29-10-2025. • Expected starting date: 1st of December 2025. • All candidates who formalize their applications in an improper way or fail to prove the requirements imposed by this tender are excluded from admission. In case of doubt, the panel is entitled to request any candidate to present further documentation supporting their statements. • False statements provided by the candidates shall be punished by law. • The list of both admitted and excluded candidates and final classification list shall be notified by email. • After notification, all candidates have 10 working days to respond. Panel’s final decisions are pronounced within a period of 90 working days, from response period deadline. • This tender is exclusively destined to fill this specified position and can be terminated at any time until approval of final candidate list, expiring with the respective occupation of said position. • Non-discrimination and equal access policy: Associação Fraunhofer Portugal Research actively promotes a non-discrimination and equal access policy, wherefore no candidate can be privileged, benefited, impaired or deprived of any rights whatsoever, or be exempt of any duties based on their ancestry, age, sex, sexual preference, marital status, family and economic conditions, instruction, origin or social conditions, genetic heritage, reduced work capacity, disability, chronic illness, nationality, ethnic origin or race, origin territory, language, religion, political or ideological convictions and union membership. • The panel has approved this announcement in meeting held on 15/10/2025. • To comply with Law no. 4/2019, of 10 January, candidates must declare on the application form, under a statement of honor, their degree of disability, the type of disability and the means of communication/expression to be used in the selection process.