HUN-REN and Óbuda University launch agent-based AI program to accelerate research across Hungary
Date: 14.01.2026
HUN-REN and Óbuda University have formed a long-term consortium to build AI-driven tools that speed up academic research and turn ideas into real outcomes. The plan centers on an Agentic Discovery Platform-AI agents that automate specific steps in the research lifecycle, from hypothesis generation to publishing.
The goal is simple: make AI a default part of researchers' daily work, not an add-on. Beyond the platform, the partnership will provide training, open-source tools, and an AI service center so teams can adopt proven methods fast.
What's being built
- Agentic Discovery Platform: autonomous AI agents supporting literature triage, experimental design, data analysis, and paper drafting.
- AI services for labs and institutes: consulting, setup, integrations, and support.
- Training and an AI ambassador program: hands-on enablement inside departments to drive adoption.
- Open-source tooling: reusable components that standardize workflows and promote reproducibility.
If you want a primer on LLM-based autonomous agents, this survey is a solid starting point: A Survey on LLM-based Autonomous Agents (arXiv).
Why it matters for engineers and data teams
- Standardized research pipelines: shared components for data prep, retrieval, evaluation suites, and reporting.
- API-first agents: orchestrate tools, search, code execution, and data access under clear policies and quotas.
- Model operations: registries, prompt/version management, offline and online evaluation, rollback paths.
- Secure deployment patterns: on-prem, private cloud, or hybrid with role-based access, audit logs, and data isolation.
- Interoperability: connectors for publications, knowledge graphs, vector stores, ELN/LIMS, and HPC queues.
Expected impact
- Weeks shaved off literature reviews via targeted retrieval and summarization.
- Faster method iteration with templated experiment planners and checklists.
- Higher throughput from automation of repetitive analysis and formatting tasks.
- Better reproducibility with versioned data, prompts, and notebooks tied to each result.
Governance and compliance
The consortium plans to bake governance into workflows: data classification, consent tracking, model cards, risk controls, and audit trails. For context on regulatory direction in Europe, see the EU AI Act overview.
Rollout and enablement
A central service center and an AI ambassador program will help teams pick the right tools and ship working solutions. Expect a mix of platform services, training, and open-source releases so institutes can adopt at their own pace.
If your team needs structured upskilling, this curated index can help: AI courses by job role.
Leaders' perspective
Balázs Gulyás, President of HUN-REN, set a clear objective: make Hungary an active participant in the global AI push. He also pointed to tighter links between university research and HUN-REN institutes, especially by engaging students early.
Roland Jakab, CEO of HUN-REN, highlighted tools that automate and speed up value-creating work. "We have set up a service centre and an AI ambassador programme to help researchers identify the tools and methods best suited to their scientific work," he said.
Levente Kovács, Rector of Óbuda University, emphasized that researchers bring natural intelligence while AI opens a new dimension for R&D. Applied well, he expects research processes to run several times faster.
What comes next
The consortium is a long-term partnership aimed at measurable scientific, innovation, and economic gains. It will expand by inviting domestic and international partners with strong research capacity to join and strengthen the ecosystem.
Technical opportunities for contributors
- MLOps/LLMOps: model registries, eval suites, safety checks, and deployment toolchains.
- Data engineering: ETL for publications, datasets, lab systems, and metadata governance.
- Retrieval systems: hybrid search, knowledge graphs, and domain ontologies.
- Agent orchestration: tool calling, planning strategies, and guardrails.
- Security and compliance: access controls, isolation, logging, and policy enforcement.
Bottom line: this is a practical path to AI-assisted research at scale-built on agents, governance, and repeatable engineering.
Your membership also unlocks:
