Kroonen AI is a privacy-first AI infrastructure provider based in Delaware, USA. We offer four core services: custom AI deployment with self-hosted infrastructure, model fine-tuning for open-weight models like LLaMA, Mistral, Qwen, and OpenAI GPT-OSS, comprehensive safety evaluation including red team testing and CBRN assessments, and full-stack application development using React, TypeScript, Node.js, and iOS SwiftUI. Our focus is on data sovereignty - your data never leaves your control.

What models can Kroonen AI fine-tune?

Kroonen AI specializes in fine-tuning open-weight large language models including Meta's LLaMA family (LLaMA 2, LLaMA 3), Mistral AI models (Mistral 7B, Mixtral), Alibaba's Qwen series, and OpenAI GPT-OSS family models. We also implement memory frameworks with visual and textual embeddings for building advanced AI agents. Fine-tuning is performed on your infrastructure or secure private environments to maintain data privacy.

What AI safety evaluation services does Kroonen AI offer?

Kroonen AI provides comprehensive AI safety evaluation services including: red team testing to identify vulnerabilities and potential misuse vectors, CBRN (Chemical, Biological, Radiological, Nuclear) evaluation aligned with frontier lab standards, ASL-3 style capability assessments, benchmark analysis against standard safety metrics, and alignment testing. Our methodology follows practices established by leading AI safety organizations like Anthropic and OpenAI.

What makes Kroonen AI's infrastructure privacy-first?

Kroonen AI's privacy-first approach means: all AI infrastructure runs on-premise or in private cloud environments you control, zero telemetry - no usage data or prompts are collected or transmitted, no third-party API dependencies that could expose your data, complete data sovereignty with full audit trails, and GDPR/HIPAA-compatible deployment options. Unlike cloud AI services, your sensitive data never leaves your environment.

What technology stack does Kroonen AI use for application development?

Kroonen AI builds applications using modern, production-ready technologies: Frontend - React with TypeScript, Tailwind CSS, and Framer Motion for web applications; Backend - Node.js with Express or Fastify, Python for ML pipelines; Mobile - Native iOS development with SwiftUI; AI Integration - Ollama for local LLM deployment, custom API wrappers, and RAG (Retrieval-Augmented Generation) implementations. All applications are designed with privacy-by-default architecture.

How does Kroonen AI handle custom AI deployment?

Kroonen AI's custom deployment process includes: infrastructure assessment and planning, selection and optimization of appropriate open-weight models, deployment using Ollama or custom serving solutions, integration with existing systems via REST APIs or SDKs, monitoring and observability setup, and ongoing maintenance support. Deployments can be on bare metal, Kubernetes, or private cloud providers. We handle everything from single-GPU setups to distributed inference clusters.

Libre WebUI is an open-source project by Kroonen AI - a fully libre, privacy-respecting web interface for local LLMs. It features complete Ollama integration, zero telemetry, no cloud dependencies, keyboard shortcuts, and a minimalist design optimized for performance. It's available at librewebui.org and the source code is on GitHub at github.com/libre-webui/libre-webui.

How can I contact Kroonen AI?

You can reach Kroonen AI via email at hello@kroonen.ai for general inquiries and project discussions. For technical discussions and community support, visit our GitHub Discussions at github.com/libre-webui/libre-webui/discussions. We typically respond within 24 hours. You can also find us on X/Twitter (@kroonen_ai), GitHub (kroonen-ai), Mastodon (@kroonen_ai@mastodon.social), and Bluesky (@kroonen.ai).

Does Kroonen AI work with enterprises?

Yes, Kroonen AI works with enterprises, startups, and research organizations. Our services are designed for organizations that require: data sovereignty and regulatory compliance (GDPR, HIPAA, SOC2), custom AI solutions tailored to specific domain needs, reduced dependency on third-party cloud AI providers, in-house AI capabilities with full control over models and data, and comprehensive safety evaluation before deploying AI systems.

What differentiates Kroonen AI from cloud AI providers?

Unlike OpenAI, Anthropic, or Google Cloud AI, Kroonen AI focuses on self-hosted solutions where you maintain complete control. Key differences: your data never leaves your infrastructure, no per-token API costs after initial deployment, models can be fine-tuned specifically for your use case, no vendor lock-in with open-weight models, full transparency into model behavior and outputs, and the ability to run air-gapped for maximum security.

Kroonen AI - Privacy-First AI Infrastructure & Custom Model Fine-Tuning

Kroonen AI Kroonen AI is a privacy-first AI infrastructure provider offering custom AI deployment, model fine-tuning for LLaMA/Mistral/Qwen/GPT-OSS, safety evaluation with red team testing and CBRN assessments, and application development with React/TypeScript/Node.js/iOS. Based in Delaware, USA. Contact: hello@kroonen.ai https://www.kroonen.ai

Robin Kroonen AI Infrastructure & Services

2024 Custom AI Deployment - Self-hosted infrastructure with Ollama, LLaMA, Mistral, Qwen Model Fine-Tuning - LoRA, QLoRA, full fine-tuning for open-weight LLMs AI Safety Evaluation - Red team testing, CBRN evaluation, ASL-3 assessments Application Development - React, TypeScript, Node.js, iOS SwiftUI, privacy-first design hello@kroonen.ai Worldwide