About PKG Systems
PKG = Packaged. That's the name, and it's the whole idea.
PKG Systems has been delivering custom technology solutions since the late 1990s. Over those 25 years we've built systems for companies of every size — startups, mid-market firms, Fortune 100 corporations, state and federal government. We've watched the industry shift from on-premise data centers to managed hosting to public cloud to multi-cloud to "AI-native everything," and we've stayed in business through all of it because the underlying job hasn't changed: companies have data and processes that don't fit in a box, and they need someone competent to build the box for them.
What's changed is who we choose to work with, and how we package the work.
Mid-market, by choice
We focus on mid-market companies — roughly $20M to $500M in annual revenue, horizontal across industries. That's not a fallback position; it's a deliberate choice. We've worked at every scale, and mid-market is where the work is most rewarding for everyone involved.
- Mid-market companies need real help. Big enterprises have 500-person IT departments. Mid-market companies have a CTO, an ops manager, and a wishlist longer than either of them can get to. They actually need the thing built.
- The decisions get made by people who care about the outcome. Not by procurement committees, not by political stakeholders, not by external consultants with their own agendas. The CFO who signs the contract is the same person whose business it is to make work.
- The work is honest. When a mid-market company succeeds, you can see it. When the system goes down, you hear about it directly. The feedback loop is short and real.
- Senior engineers, mid-market scope. We can put senior people on every engagement because we're not pricing for a Fortune 500 IT budget. The work matches the skill.
What we won't do
Honesty about positioning is more useful than honesty about features. Here's what we don't do, and why.
We don't subcontract under Big 4 firms
Twenty-five years of watching the largest consulting firms collect a markup and pass the actual work down a chain. We've been the bottom of that chain. We won't do it again — at any price. If you're working with a Big 4 prime, we can't help you within that engagement.
We don't chase federal procurement
We've delivered on government contracts as a subcontractor and we know the work. We just won't take direct federal procurement engagements anymore — the cost of compliance and the political cycle aren't worth the work for a firm our size.
We don't pursue Fortune 100 accounts
Big corporations almost always default to the safe choice — the Big 4 firm, the multi-billion-dollar vendor, the "no one ever got fired for buying" option. The decision-making is usually political, the procurement gauntlet is long, and the actual work is a small fraction of the total cost. Mid-market companies pick on merit. That's where we want to be.
We don't bill time and materials
Every engagement is a packaged solution: discovery, fixed-price build, monthly run-rate. We're not paid to log hours; we're paid to deliver and run the work. If T&M is what you want, we're not the right fit, and that's fine.
We don't sell SaaS
Your data lives in a cluster dedicated to you. We don't operate a multi-tenant database with your records mixed into someone else's. We don't run an analytics engine on top of an aggregated customer dataset. The economic model is different and we like it that way.
What we do
Private AI Data Platforms — your data from every system, in a private Kubernetes environment we build for you, with AI tools your team uses to analyze it. That's the headline service today.
Custom Kubernetes environments, custom applications, and enterprise data migration are the three capabilities that make the data platform possible. They're also things we'll do as standalone engagements when that's what the customer needs.
Our approach is the same on every engagement: discovery, fixed-price build, monthly run-rate. One firm, one contract, end-to-end.
Read the case study for an anonymized walkthrough of a real engagement running in production.
How we work, technically
We have opinions and we use them consistently:
- Kubernetes for infrastructure — single-tenant clusters, customer-isolated by architecture
- Postgres for data — the most boring, most reliable choice in the industry, and still the right one
- Python (FastAPI + asyncio) for backend services and data daemons
- SolidJS for admin UIs — small, fast, modern
- Authentik for identity — OIDC throughout, role-based access on every internal tool
- Self-hosted everything — including analytics, identity, registries, and AI inference
We don't use third-party SaaS in the data path. We don't send customer data to external analytics platforms. We don't depend on services we don't control for core functionality. This isn't ideological — it's how you actually deliver privacy as architecture rather than as a privacy policy.
Where we are
Based in the United States. Working in US Central Time. Available for the kind of conversations that lead to real engagements — not for "discovery calls" that are sales pitches in disguise.