Trustworthy GenAI Evaluation for GenAI4EU Booster — HORIZON-CL4-2026-04-DIGITAL-EMERGING-19

Giskard is a French SME developing an open-source platform for evaluating and red-teaming generative AI systems. We are looking to join a consortium for HORIZON-CL4-2026-04-DIGITAL-EMERGING-19: Challenge-Driven GenAI4EU Booster in Apply AI prioritised sectors.

What we bring to a consortium:

• Automated evaluation of generative AI models: hallucination detection, factual accuracy assessment, and output quality benchmarking

• Security testing for LLM-based applications: prompt injection, data leakage, jailbreak resistance

• RAG pipeline evaluation through our RAGET toolkit (automated test generation from knowledge bases)

• AI Act conformity assessment for high-risk GenAI deployments

• Proven open-source technology adopted by the European AI community (GitHub: 4k+ stars)

Relevant sector expertise: Our evaluation tools are sector-agnostic and can be adapted to assess GenAI reliability in pharma/drug discovery (e.g., hallucination-free molecule generation, clinical text accuracy), aerospace (e.g., safety-critical AI documentation), and telecommunications (e.g., network optimization agents).

Role we are looking for: Partner for evaluation, red-teaming, benchmarking, and AI Act compliance workpackages.

What we are looking for: A coordinator with deep domain expertise in one of the three target sectors (pharma, aerospace, telecom) and experience managing large-scale Horizon Europe RIA projects.