# NeuraNidus — Privacy-Preserving Federated Edge Intelligence
> Open-core federated learning orchestration platform. Rust. Apache 2.0. Built in Germany.

## About
NeuraNidus UG (Duisburg, Germany) builds Nidus Synchro, an open-core federated learning orchestration platform that lets organizations train ML models across distributed devices — from cloud servers to bare-metal microcontrollers — without centralizing sensitive data. The platform spans three planes (Node, Adapter, Control), supports pluggable FL frameworks (Flower, TFF, FATE), and ships with TPM hardware attestation, differential privacy, and RL-based fleet scheduling.

## Products

- [Nidus Synchro](https://neuranidus.com/products): Open-core federated learning orchestration. Rust. Apache 2.0. Shipping Q1 2027. Three-plane architecture: Node (bare-metal/RTOS agent), Adapter (framework-agnostic bridge), Control (RL-powered fleet scheduler).
- [Nidus Nexus](https://neuranidus.com/products): Dynamics 365 BC ↔ SAP Ariba procurement connector. Browser-local semantic mapping with ONNX Runtime Web and WASM fallback. AL extension, AppSource-ready. Available Q3 2026.
- [Nidus-Micro](https://github.com/NNidusLabs/nidus-micro): Bare-metal device agent. <1 MB. no_std Rust. Microcontrollers and RTOS targets.
- [Nidus-Pro](https://github.com/NNidusLabs/nidus-pro): Linux SBCs, Jetson, WebNN. Full-featured edge gateway agent.
- [Nidus-Titan](https://neuranidus.com/products): Hardened gateway with discrete TPM. Defense/industrial assurance profiles.
- [Nidus-Local](https://neuranidus.com/products): Enterprise workstation. 128–256 GB unified memory. Local LLM workloads.

## Solutions

- [Healthcare](https://neuranidus.com/solutions): Privacy-preserving diagnostics trained across hospital networks.
- [Industrial IoT](https://neuranidus.com/solutions): Predictive maintenance on distributed sensor fleets.
- [Supply Chain & Procurement](https://neuranidus.com/solutions): Semantic mapping for BC ↔ SAP Ariba. Governed FL with tamper-evident audit trails.
- [Financial Services](https://neuranidus.com/solutions): Fraud detection across distributed branches with differential privacy.
- [Public Sector](https://neuranidus.com/solutions): Privacy-preserving analytics across government agencies.
- [Energy & Utilities](https://neuranidus.com/solutions): Distributed model training across smart grid sensors.

## Research

- [Nidus Synchro White Paper](https://neuranidus.com/research): A Federated Edge Orchestrator for Privacy-Preserving Model Training. Three-plane architecture, multi-agent RL fleet scheduling, hardware-attested privacy guarantees. Coming Q1 2027.

## Engineering Principles
1. Functional Necessity — every component earns its place.
2. Modular Boundaries — standardized interfaces with strict contracts.
3. Platform Honesty — respect real constraints, no leaky abstractions.
4. End-to-End Coherence — full operational path from ingestion to outcome.
5. Error Integrity — explicit, bounded, observable failures.
6. Standardized Delivery — consistent, verifiable, containerized.

## Company

- [Company page](https://neuranidus.com/company): Founded by Anton Kornilevsky. Based in Duisburg, Germany.
- [Contact](https://neuranidus.com/contact): Enterprise inquiries, partnerships, open source collaboration.
- [GitHub](https://github.com/NNidusLabs): Open source repositories under Apache 2.0.
- [Legal Notice](https://neuranidus.com): Imprint per German law (Impressum).

## Discovery

- Sitemap: [sitemap.xml](https://neuranidus.com/sitemap.xml)
- Full site content: [llms-full.txt](https://neuranidus.com/llms-full.txt)

## Contact
- Email: contact@neuranidus.com
- Phone: +49 (176) 63074072
- Address: NeuraNidus UG, Wildstr. 26, 47057 Duisburg, Germany