ITEA 4 R&D SME PROFILE // DELFT – THE HAGUE TECHNOLOGY CORRIDOR, NL

Kleynenberg Software Ontwikkeling

Independent Dutch R&D SME led by Founder & Principal AI Architect Steven Kleynenberg, bringing over 30 years of hands-on research and engineering experience in Artificial Intelligence, massively parallel computing, Natural Language Processing (NLP), and distributed enterprise and clinical architectures.

ENTITY: SME (THE NETHERLANDS) DEPARTMENT: R&D (AI & SYSTEMS ARCHITECTURE) LEAD: STEVEN KLEYNENBERG CANONICAL: https://kso.kleynenberg.com/

01 // Core AI & Technical Expertise

Pillar 01 // Deterministic NLP + Vector Retrieval WO 2014/003543 A1

Sovereign AI & Hybrid Enterprise-RAG

Architecting privacy-by-design, on-premise, and zero-lock-in AI retrieval pipelines. As co-inventor of PCT patent application WO 2014/003543 A1 (Part-of-Speech linguistic query generation and document clustering), KSO combines deterministic linguistic pre-filtering alongside dense vector search, semantic indexing, and local Large Language Models (LLMs) to maximize retrieval precision, domain grounding, and full auditability.

Pillar 02 // Client-Side Parallelism WebGPU / WGSL / MetaNet Lineage

Browser-Native AI, WebGPU & High-Performance Compute

Building client-side, zero-footprint AI inference and parallel compute kernels using WebGPU/WGSL shaders, WebGL, and Chrome Extensions (including hardware-accelerated local speech synthesis/recognition and interactive visual neural network simulators). This work continues a research lineage dating back to modular neural network compilers and massively parallel neurocomputers (MetaNet, BSP400, and Connection Machine architectures).

Pillar 03 // Biometric & Environmental Edge BLE / LoRaWAN / 4D Analytics

Real-Time Sensor Telemetry, Edge-IoT & Decision Support (DSS)

Developing end-to-end physiological and environmental telemetry pipelines—integrating Bluetooth LE biometric wearables (Zephyr BioHarness, Polar H7), LoRaWAN/TTN edge nodes, and low-latency Linux signal/audio DSP—coupled with rule-based Decision Support Systems (DSS) and interactive multidimensional visual analytics.

Pillar 04 // Zero-Bloat Systems Engineering Linux / C / C++ / Python / Node.js

Mission-Critical Systems & Sovereign Infrastructure

Deep systems engineering on Linux (C/C++, Python, Node.js, Apache modules, and bare-metal/embedded architectures), deterministic control models, strict digital compartmentalization, and automated forensic and security infrastructure without proprietary cloud dependencies.

02 // European Public-Private Consortium Track Record

ITEA 2 · ITEA 3 · EU FP5 Architecture & Demonstrators

Proven Architecture & Work-Package Contributions

Extensive track record in European public-private innovation consortia, having contributed reference models, distributed system architectures, and working software demonstrators across five flagship programs:

  • PARTNER (ITEA 3 16017): Award-winning European collaborative innovation project on patient-centric clinical decision support and integrated care workflows.
  • Medolution (ITEA 3 14003): Next-generation medical resolution and connected clinical telemetry architectures.
  • MEDUSA (ITEA 2 10004): Medical Distributed Utilization of Services & Applications for time-critical acute stroke workflows (published at ESOCC 2015).
  • EDAFMIS (ITEA 2 07011): Enterprise Data & Federated Management Information Systems.
  • LORE (EU FP5 IST-1999-12645): Knowledge management and linguistic retrieval architectures.

03 // ITEA 4 Consortium Ambition & Collaboration Scope

Open for Consortium Formation ITEA 4 Calls

Partnering with European Industry, Hospitals & Research Institutes

Through Kleynenberg Software Ontwikkeling (a nimble Dutch R&D SME based in the Delft/The Hague technology corridor), Steven Kleynenberg actively collaborates with European industrial partners, clinical centers, and academic institutes on:

  • Sovereign & Trustworthy AI: Auditable, on-premise AI architectures compliant with European data sovereignty requirements.
  • Edge & Browser-Native Intelligence: Zero-footprint WebGPU/WGSL inference and low-power edge sensor processing.
  • Hybrid RAG for Knowledge-Intensive Domains: Combining linguistic ontology/POS structures with vector embeddings for patent, legal, medical, and engineering corpora.
  • Human-in-the-Loop Clinical & Operational Decision Support: Real-time telemetry fusion and transparent decision support interfaces.

Consortium Roles: System & AI Architecture Design, Work-Package Leadership, and Rapid Software Demonstrator Development.