Expert-led infrastructure for scientific workflows.

in4r is led by Ivo Djidrovski, PhD, combining toxicology and safety-science workflow experience with practical AI infrastructure development.

Portrait of Ivo Djidrovski
Founder

Ivo Djidrovski, PhD

Built by someone who has lived inside the scientific workflow.

in4r is led by Ivo Djidrovski, a scientist and AI systems builder working at the intersection of toxicology, computational workflows, and auditable agent infrastructure. The work is practical: make AI useful where evidence, data privacy, and expert accountability matter.

Scientific foundation

PhD and postdoctoral work across stem-cell biology, toxicology, and next-generation safety assessment.

AI infrastructure builder

Creator/maintainer of public open-source tooling, including O-QT-related workflows and ToxMCP modules.

Published and citable

O-QT publication, ECETOC AI trust work, public VHP4Safety-related outputs, and ToxMCP preprint/suite development.

Community validation

ONTOX and MIT/Roche hackathon wins plus practical work with research consortia and safety-science teams.

Peer-Reviewed Science Behind the Tools

Our infrastructure is backed by published research in computational toxicology. Every tool we build has a scientific foundation you can cite.

ToxMCP: Guardrailed, Auditable Agentic Workflows for Computational Toxicology via the Model Context Protocol

PREPRINT
Ivo Djidrovski·bioRxiv·2026·DOI: 10.64898/2026.02.06.703989
MCP ProtocolOpen SourceSafety ScienceToxMCP

User-driven development and evaluation of an agentic framework for analysis of large pathway diagrams

PUBLISHED
Corradi, M., Djidrovski, I. et al.·Bioinformatics Advances·2026·DOI: 10.1093/bioadv/vbag277
Agentic AIPathway DiagramsUser Evaluation

Building trust in the integration of artificial intelligence into chemical risk assessment: findings from the 2024 ECETOC workshop

Gant, T.W. et al. including Djidrovski, I.·Archives of Toxicology·2026·DOI: 10.1007/s00204-025-04286-8
AI TrustEvidence GovernanceECETOC

O-QT assistant: A multi-agent AI system for streamlined chemical hazard assessment and read-across analysis using the OECD QSAR toolbox API

PUBLISHED
Ivo Djidrovski, Raymond Pieters, Juliette Legler, Marc Teunis·Computational Toxicology·2025·DOI: 10.1016/j.comtox.2025.100395
Multi-agentOECD QSAR ToolboxRead-acrossO-QT

The Virtual Human Platform for Safety Assessment (VHP4Safety) project: Next generation chemical safety assessment based on human data

Kienhuis, A.S. et al.·ALTEX·2025·DOI: 10.14573/altex.2407211
VHP4SafetyHuman DataNew Approach MethodologiesVHP4Safety

Report of the First ONTOX Hackathon

PUBLISHED
Diemar, M.G. et al.·Alternatives to Laboratory Animals·2025·DOI: 10.1177/02611929241305112
ONTOXHackathon WinnersWorkflow Prototyping

Where the Work Is Presented

Public talks, workshops and posters on auditable AI for safety science.

12 Nov 2026·NCAC-SOT · Virtual

ToxMCP: Connecting AI Agents to Toxicology Tools and Evidence

UPCOMING

Invited talk.

6 Nov 2026·SST Annual Meeting · Basel

Applying Artificial Intelligence to Gather Toxicological Data

UPCOMING

Hands-on ToxMCP workshop, co-led with Joyce de Paula Souza (SCAHT). Fully booked.

Talk and workshop resources
5 Nov 2026·SST Annual Meeting · Basel

Building auditable human controlled AI systems

UPCOMING

Invited talk, session "From Data to Insight: AI & Digital Toxicology".

Talk and workshop resources
13–16 Sep 2026·EUROTOX 2026 · Vienna

ToxMCP: Auditable, Reproducible Workflows for NAM-Based Toxicology and Risk Assessment

PAST

Poster P30-36, and late-breaking poster LP-45 on ToxMCP-enabled ontology grounding of AOP-Wiki with J. Filipovska.

Jun–Jul 2026·ESTIV 2026 · Maastricht

ToxMCP and O-QT posters (P55, P56)

PAST

Guardrailed, auditable agentic workflows via the Model Context Protocol, and auditable AI-assisted QSAR Toolbox read-across workflows. With Juliette Legler, Raymond Pieters and Marc Teunis.

Poster resources
Sep 2025·EUROTOX 2025

Augmenting Chemical Hazard and Risk Assessment through Large Language Models and AI Agents

PAST

Poster P03-13, with Marc Teunis.

Start with one case-study pilot or advisory retainer.

Bring one safety-science or research case study. We will define the pilot boundary, consulting scope, evidence sources, review gates, and deployment path before scaling anything. We are opening pilot and design-partner conversations for teams operationalizing AI in review-heavy scientific workflows.

01Case study
02Consulting scope
03Build path
What you get from a pilot
Pilot protocol
Working prototype
Evidence bundle: PDF, tables, figures
Audit-log example
Uncertainty + failure-mode readout
Build/governance roadmap
Book a pilot scoping call

Tell us what case study, workshop, or workflow you want to improve.

Private inbox

Please do not submit confidential data through this form. We use your details only to respond to your inquiry. Workflow details are treated confidentially. See the privacy note.