OptimAIse

Artificial Intelligence (AI) and Generative AI (GenAI) development faces critical Software Engineering challenges (e.g., performance, modularity, automation, trustworthiness…) that hinder EU SMEs and small research groups from taking the global lead and threaten EU sovereignty. OptimAIse addresses these challenges by delivering a scalable, modular, and interoperable reference architecture that leverages European hardware to enable efficient and simplified large language model deployments. A flexible framework – aimed to deliver streamlined operations and greater productivity for developers – combines an adaptive orchestration engine (continuous performance optimisation with energy and sustainability guardrails) with tools embedding explainability, robustness and fairness as core artefacts. Besides, an end-to-end cryptographically secure traceability infrastructure ensures transparency, reproducibility, and compliance with the EU AI Act. The OptimAIse approach will be validated in relevant environment (TRL6) over industrial use cases, showcasing specific challenges for GenAI in healthcare and robotics, also involving an AI Factory within the development and validation processes. Led by key actors in the field, a thorough evaluation using qualitative and quantitative indicators through 2 rounds of pilots will allow adjustments and refinement towards improvement, further adoption, scalation, and re-adaptation to different scenarios, opening tangible exploitation pathways. A multi-disciplinary consortium of 7 SMEs – including a leading accelerator network – 3 research centres, 1 university, and 2 public administrations join forces, leveraging their expertise and experience in past and ongoing projects – also boosting synergies in dissemination, communication and exploitation actions – to ensure the successful achievement of the objectives, positioning Europe at the forefront of high-performance and trustworthy GenAI.