About Us
Years shipping production software
Client problems solved
Industries served
Founded in Chicago, shipping since
How we got here
From building software the slow way for two decades to leading the spec-driven approach for AI-augmented engineering. Same discipline. Faster outputs.
The first two decades
LoadSys was founded in Chicago in 2004 to ship production software for businesses that needed real systems, not prototypes. 400+ client engagements across manufacturing, healthcare, education, e-commerce, professional services, and financial operations. Stanford. Mid-market companies. Mom-and-pop shops that needed the same engineering discipline as the Fortune 500 — and got it.
When AI coding agents arrived
By 2024, AI coding agents — Claude Code, Cursor, Copilot — were everywhere. So were the failures. Teams were generating mountains of plausible-looking code that fell apart in production. Most “AI-assisted development” was vibe coding at scale, with the same root cause every time: the agent had no plan to execute against.
What LoadSys is now
An AI engineering consultancy built on Spec-Driven Development. We rebuild internal custom applications in 90 days. We train engineering teams on AI coding agents. We design and deploy MCP servers, AI integrations, production agents. We built Brunel — the planning platform we use on our own work — and we ship it to other teams that need the same discipline.
Three things that shape every engagement
The discipline didn’t change when AI arrived. The tools changed. The principles didn’t.
The spec is the contract
Plan, then execute, then verify. Whether the engineer is human or an AI coding agent, the written spec is the contract between intent and output. Agile changes the spec mid-flight; waterfall writes a 200-page spec nobody reads. Spec-Driven Development writes a spec the team executes against and the agent can be measured against.
Senior review is non-negotiable
AI coding agents are fast, not wise. Every commit goes through senior-engineer review before merge — same as any serious shop. The agent doesn’t ship unsupervised. The engineer is the source of truth. Speed without supervision is how teams accumulate AI-generated technical debt they can’t see.
If we can’t ship it, we say so
We walk away from projects we can’t deliver. The Discovery Sprint exists so that the “no” comes before the contract, not after. Our 90-day modernization offer puts 25% of the fee at stake if we miss the deadline — which means we scope tightly or we don’t take the engagement.
Brunel — the planning layer your AI agents were missing.
82% of AI agent failures trace to poor pre-execution planning. Brunel sits between dev teams and Cursor, Claude Code, Copilot, providing the Plan → Export → Execute → Verify workflow that closes the gap.
We built Brunel because we needed it. We ship it to other teams who do too. The completion illusion — when an agent reports “done” but only 30–40% of the work is finished — disappears when the spec is the contract.
04 · Leadership
The team that ships
Practitioner-led. Both founders write code, write specs, and run engagements. We don’t have an account-management layer between you and the engineers.
Lee Forkenbrock
CEO · ChicagoI’ve been shipping production software since the early 2000s. I lead LoadSys’s engagement strategy: which projects we take, which we walk away from, and how we scope the ones we sign.
Spec-Driven Development is the discipline I’ve watched separate teams that ship from teams that stall for two decades — long before AI agents made it urgent. The methodology didn’t change when the tools did. The stakes did.
On any given engagement, I’m the one who calls when a project isn’t going to ship in 90 days — and the one who says “yes, we can scope this” when it can.
Donatas Kairys
President & CTO · ChicagoI architect LoadSys’s AI engineering practice and lead Brunel’s product direction. Nineteen years of building data-intensive systems before AI — cloud infrastructure, Databricks implementations, large-scale data analytics — and now the work is teaching AI coding agents to ship against a spec.
I’m the one designing the engagement architecture on most modernization projects, from the spec to the production cutover. If you’ve worked with us, you’ve probably been on a call with me.
Outside of client work, I’m deep in the Anthropic ecosystem — Claude API, MCP, agent architecture — because the practitioners we recommend to our clients should be the practitioners we are.
Six industries, one methodology
The integration patterns differ by industry. The discipline of shipping to production doesn’t.
Manufacturing
MES systems, production scheduling, operational dashboards. The internal-app modernization sweet spot.
Healthcare
Internal admin tools, operational workflows, document processing where compliance matters.
Education
Operational portals, administrative systems, vertical SaaS for institutions. Stanford among the named references.
E-commerce
Admin tools, vendor portals, integration layers, custom merchant systems.
Professional Services
Internal applications for law, accounting, consulting. Document-heavy workflows where LLM integration pays back fast.
Financial Operations
Internal admin layers, decision-support tooling, integration with established financial systems.
Next step
Talk to a senior AI engineer.
30 minutes with Lee or Donatas — not an account manager. We’ll discuss your project, give you a straight answer on fit, and recommend a different path if LoadSys isn’t right for the work.