Your team is already using AI tools. Do you know where your data is going?
I build private AI systems for businesses that can't afford a breach -- and can't afford to fall behind. Healthcare and legal firms trust me because the stakes there are highest. The problem isn't unique to them.
Most SMBs are adopting AI tools without a plan, a risk assessment, or anyone accountable when something breaks. That's what I fix.
Your employees are already using AI. The question is whether your data stays in the building.
Most SMBs don't find out their IT is broken until a client does.
Most SMBs know they need AI. Few know where to start — or what to avoid.
Real systems, built for real constraints. Every project here is designed around the security and compliance needs of the clients who can't afford to get it wrong.
Your documents, answered instantly — without leaving your network.
A private RAG-based retrieval system built for healthcare, legal, and professional services firms. Staff get instant answers from internal documents without exposing sensitive data to any third-party AI service. Deployed on dedicated hardware using vLLM for inference, Open WebUI for the interface, and Qdrant for vector storage. Designed for regulated industries where data sovereignty isn't optional — your documents never leave your network, ever.
Proof that a private LLM can outperform a general one — on your terms.
A working demonstration of domain-specific AI: a hybrid chess engine combining Qwen 2.5 7B with Stockfish and Lc0, deployed on Proxmox with dedicated GPU inference. The same architecture applies to any specialized knowledge domain — legal research, clinical protocols, internal policy. If you can define the domain, a private model can be trained to master it.
Enterprise-grade AI infrastructure — sized and priced for SMBs.
A fully documented, deployable architecture for organizations that need private AI without enterprise budgets. Built and validated on a 12-node lab: Proxmox hypervisor, dedicated GPU inference, vLLM serving with Open WebUI and Qdrant vector storage, VLAN-segmented networking, and LibreNMS monitoring. Everything needed to run a production private AI stack — designed to be handed off and maintained by your existing IT team.
From invoice to reconciliation — automated where it counts.
An AI-assisted workflow designed to eliminate manual billing overhead for SMBs. Will integrate with QuickBooks and other platforms to automate invoice generation, payment tracking, and reconciliation — adapted to each client's existing process, not the other way around.
Most IT consultants learned their trade on SMB infrastructure. I learned mine at Toyota North America, TV Guide, and Fox Sports -- environments where downtime had names attached to it and decisions had real consequences.
That experience includes serving as Project Manager for the Toyota North America data center consolidation, leading the infrastructure build for TV Guide's Los Angeles data center, and running MIS operations at Fox Sports. These aren't resume line items -- they're the reason I know exactly which decisions break systems and which ones don't.
Now I apply that same rigor to small and mid-sized businesses -- organizations that deserve enterprise-grade thinking without the enterprise overhead. Whether it's private AI infrastructure, managed IT, or helping leadership make their first serious technology decision, the standard doesn't change.
Practical AI and IT content for operators who want to understand the tools -- not just use them. Built from a real home lab, for real-world problems.
Subscribe on YouTube
The background behind this channel -- 30+ years across Toyota, UCLA Health, and MSP operations. Why that experience changes how you should think about AI in your business.
Frontier AI tools are convenient. They're also quietly creating liability for businesses that use them. What your legal and IT teams need to understand before the next breach.
One HIPAA network breach. One financial fraud forensics job. What actually happens when something goes wrong -- and what most businesses get wrong before it does.
Most businesses don't choose how they adopt AI -- they drift into it. Here's a framework for understanding the three levels, and why most businesses are stuck at the wrong one.
Read post →"Clear communication, predictable delivery of our infrastructure. Chris and his team relocated our Torrance office and upgraded all three of our office locations, including Beverly Hills."
"IT CloudLink modernized our systems without disrupting the offices. They also quickly remediated a cyber attack on our offices. Security-first thinking and dependable technical support."
"ITCL is reliable, responsive, and gives practical guidance on technology decisions. Clear advice. They're the best!"
"I don't know what we'd do without Chris. In addition to support, he also provides cybersecurity forensics and expert witness services. I depend on him daily."
"Mr. Rondthaler maintains a high level of expertise in a wide variety of disciplines. I would recommend him for any job that requires a high degree of technical knowledge."
"I worked with Chris on a project building a new technology infrastructure. Chris was competent and reliable in providing IT infrastructure support from the client side of the project."
"Chris and I worked together on infrastructure projects, including the technical redesign and migration of corporate, IT and broadcast production to new facilities."
Have a project in mind or just want to explore what's possible? Reach out and I'll get back to you within one business day.
Or email directly: letstalk@rondthaler.dev
