EXPERANOS | AI Research & Engineering Partner for Horizon Europe | Privacy-Preserving AI, Edge Vision, Robotics

IMAGE RECOGNITIONOBJECT RECOGNITIONFEDERATED LEARNING ZERO-KNOWLEDGE MLDIGITAL TWINSLLM FINE-TUNING RETRIEVAL-AUGMENTED GENERATIONEDGE AIROS 2 ROBOTICS HOMOMORPHIC ENCRYPTIONPOSE ESTIMATIONPREDICTIVE MAINTENANCE IMAGE RECOGNITIONOBJECT RECOGNITIONFEDERATED LEARNING ZERO-KNOWLEDGE MLDIGITAL TWINSLLM FINE-TUNING RETRIEVAL-AUGMENTED GENERATIONEDGE AIROS 2 ROBOTICS HOMOMORPHIC ENCRYPTIONPOSE ESTIMATIONPREDICTIVE MAINTENANCE

TRL 4 → 8

We industrialise research: from lab-validated prototypes to operational pilots

6 domains

Privacy-preserving AI · Vision · Robotics · GenAI · Digital twins · MLOps

8 WP roles

From core AI development to technical coordination and pilot leadership

48 h

Typical response time to consortium invitations and expressions of interest

Capabilities

Deep technical expertise, engineered for deployment

Six complementary domains. In a consortium or a client project we bring them as concrete, containerised, documented software components, not slideware.

Privacy-Preserving AI & Cryptography

Federated learning across institutions, fully homomorphic encryption (FHE), zero-knowledge machine learning (ZKML), secure multi-party computation and differential privacy - mathematical guarantees for GDPR-critical data.

Flower / FLFHEzk-SNARKs

Computer Vision & Edge AI

Real-time image recognition, object recognition, tracking and pose estimation optimised for the edge: NVIDIA Jetson, TensorRT quantisation and pruning for low-latency inference on industrial and infrastructure deployments.

YOLOTensorRTJetson

Generative AI & LLM Systems

Sovereign retrieval-augmented generation (RAG), domain-specific fine-tuning (LoRA, DPO) of open-weight models, and multi-agent systems with human-in-the-loop safety, deployed on-premises or on EU cloud.

RAGLoRA / DPOAgents

Robotics & Autonomous Systems

ROS 2 navigation and perception stacks for autonomous mobile robots, robotic manipulation and collaborative cells, validated in simulation, then demonstrated in operational environments.

ROS 2AMR / cobotsSim2Real

Digital Twins & Predictive Maintenance

Sensor-fused virtual replicas of physical assets for manufacturing, energy and buildings, with predictive models that forecast failures and optimise energy use in real time.

IoT fusionForecastingSimulation

Trustworthy AI & MLOps

Explainability, robustness testing, model documentation and monitoring pipelines aligned with the EU AI Act, so consortium outputs are auditable, reproducible and exploitable.

XAIEU AI ActCI/CD for ML

The TRL bridge

We close the gap between the lab and the pilot site

Coordinators bring us in for the hardest part of a project: turning validated research into robust systems that survive contact with real operational environments.

TRL 1–3

Basic research

Principles observed, concept formulated, the domain of our academic partners.

TRL 4

Lab validation

We harden research code into engineered, testable components.

TRL 5–6

Relevant environment

Integration, benchmarking and demonstration under realistic conditions.

TRL 7

Operational prototype

Pilot deployment at end-user sites with measurable KPIs.

TRL 8

System complete

Qualified system, documentation and exploitation-ready packaging.

Interactive lab · running live in your browser

Don't take our word for it. Watch it compute.

Four of our core systems, miniaturised as live widgets: a 3D robotic digital twin with real forward kinematics, a sovereign RAG assistant, a decision-support forecasting console and a blockchain-anchored AI audit trail.

Under the hood

The stack, end to end

From perception at the edge to privacy-preserving training and verifiable inference: one pipeline, engineered as containerised components. The code below is an illustrative view of how our deployment stack composes.

federated_pilot.py

# EXPERANOS deployment stack (illustrative)
from experanos.vision import EdgeDetector
from experanos.fl import FederatedTrainer

# 1 · Edge perception: object recognition on-device
detector = EdgeDetector(
    model="yolov8s", precision="int8",
    device="jetson-orin",          # 60 FPS, <12 ms
)

# 2 · Privacy: federated training, data never moves
trainer = FederatedTrainer(
    sites=["site-a", "site-b", "site-c"],
    aggregation="fedavg", dp_epsilon=2.0,
)
trainer.round(model=detector.backbone)

# 3 · Trust: verifiable inference for the audit trail
result = detector.infer(frame, prove=True)
assert result.proof.verify()       # zk-SNARK ✓

Reference pipeline: edge perception → federated training → registry → twin/ops → OTA

For industry clients

Not writing a proposal? Hire us directly.

The same engineers who build EU research pilots deliver production AI systems for industry: from custom models and dashboards to decision-support systems, AR/VR and blockchain-backed trust infrastructure.

Generative AI & assistants

Sovereign RAG, fine-tuned LLMs and agent workflows on your infrastructure.

Dashboards & decision support

Executive dashboards, DSS, forecasting and what-if simulators your team acts on.

Vision & robotics integration

Image and object recognition on the line, robotic cells and AMR deployments.

Digital twins & AR/VR

Live 3D twins, predictive maintenance and immersive training environments.

Blockchain & trusted data

Tamper-evident audit trails and data-integrity anchoring for regulated AI.

MLOps, cloud & compliance

CI/CD for ML, monitoring, GDPR and EU AI Act compliance engineering.

Explore all services

Funding programmes

Where we contribute

We join proposals as a technology-provider SME, pilot integrator or work-package leader across the main European R&I instruments.

Illustration of AI, robotics and edge devices connected across a map of Europe

Horizon Europe

Research & Innovation Actions (RIA) and Innovation Actions (IA), with particular fit in:

Cluster 3, Civil Security for Society (disaster resilience, infrastructure protection, border & cybersecurity)
Cluster 4, Digital, Industry & Space (AI, data & robotics partnerships, trustworthy AI, advanced manufacturing)
Cluster 5, Climate, Energy & Mobility (smart grids, predictive maintenance, CCAM)
Cluster 6, Food & Natural Resources (precision agriculture, environmental monitoring)
Cluster 1, Health (privacy-preserving analytics on clinical data)

Beyond Horizon Europe

We also participate in, and monitor open calls under:

EIC Pathfinder & Transition, deep-tech maturation of cryptographic AI
Digital Europe Programme, AI adoption, data spaces, cybersecurity capacity
Cascade funding / FSTP open calls from ongoing EU projects
Eurostars & national bilateral R&D schemes

Currently open to consortium invitations for upcoming Horizon Europe Work Programme deadlines.

For proposal writers

Everything you need to add us to Part A and Part B

Writing a proposal is hard enough. Our partner page gives you ready-to-paste participant descriptions, our PIC and legal data, typical work-package roles, capability keywords and a downloadable partner profile - so including EXPERANOS takes minutes, not emails.

Open the partner kit Partner profile (PDF)

Legal name	EXPERANOS
PIC	885333163
Entity type	Private for-profit SME (validated)
Country	Athens, Greece (EU)
Role types	Technology provider · Pilot integrator · WP leader
Contact	info@experanos.eu

Evidence, not promises

See the technology running

Short demonstrations from our engineering stack. Live or on-site demos available for consortium meetings and proposal preparation.

Real YOLO footage: people detected and tracked with bounding boxes and IDs

Real-time object detection & tracking

Live YOLO inference: every person detected, tracked and counted, frame by frame.

Real footage · Ultralytics YOLO

Human pose estimation

133-keypoint whole-body tracking (body, face, hands) across multiple people at once.

Real footage · OpenMMLab MMPose

Real robot arms performing learned manipulation tasks

Robotic manipulation

Learning-based robotic control: real arms executing vision-guided manipulation.

Real footage · Hugging Face LeRobot

All demonstrations

Applied AI engineering for
European research consortia