Portland, Oregon
Custom software,
built fast,
kept running.
One person accountable. Software that wasn't economically possible a year ago is now possible for your business. I'm Landon — 25 years building software, now working directly with small businesses.
If any of this is your week
Most people don't arrive describing the software they want. They arrive describing a Tuesday — and a quiet vision of what could be if just the right piece of software existed for them. Take both seriously.
Music school director
I spend Sunday nights with a spreadsheet and graph paper. The technology I'm using is old — there's got to be a better way. I'd give myself and the parents one shared calendar, in an app they actually want to use. It could be really simple. I just need someone to help me build it.
Scheduling students against teachers, rooms, instruments, and family constraints is a real computational problem. Software like this used to live only inside big logistics operations. It can be built for a music school now.
Food cart pod operator
I run a food cart pod. Twenty carts on site, each running their own order system, all of them serving the same beverages. Customers don't experience us as a place — they experience twenty separate transactions. I want a menu and order system built for my unique business. I know exactly what I need. I just need someone to help me build it.
The pod is already a brand. Right now it's a brand without a portal — and for any customer who isn't standing on the gravel, the portal is the brand. The system that turns twenty vendors into one place is buildable now.
Independent book author
I make books with contributors, indexes, attributions, copyrights on every page. Every release is the same shape: organize the assets and metadata, then publish manually in triplicate — to the printer, to the website, to the ebook store. Same three steps, a thousand times. What I want is an AI in the cloud that keeps everything organized and runs my whole operation. I know it can be done. I know the time is right.
The release cycle has the same shape every time. Software that knows the shape can run most of it — assets, contributors, copyrights, distribution channels, the catalog as it grows. The scale that used to justify that kind of system was much larger than yours. That isn't true anymore.
None of these is the question “I need custom software.” The question is whether you want to run your small business with the precision and speed of an enterprise system — built for your business, shaped to your vision of what's possible. If you recognize yourself in any of this, here's a longer read on what kind of software does what →
What just changed
The kind of software the giants spent twenty years building — sized down for the long tail.
The Ubers and Chipotles, the Nikes, the Deltas, the Intels, the Qualcomms — these enterprises run on custom software that matches exactly their business, their workflows, their customers, their brand, their data. Built with the same tools and technologies we’ve been using for twenty, twenty-five, thirty years. That kind of system has never been available to you, your business, your customers, your employees.
What changed is the technology. Large language models, coupled with coding tools and software platforms that have finally abstracted software into reasonable building blocks — capable in their own right, and far more capable in the hands of a capable practitioner. What used to take years now takes weeks. What used to require a team can be done by one person. The cost has dropped by multiple orders of magnitude.
The barrier was never whether the problem was real. Small operators have always had the same shape of problem as the giants. The barrier was cost. The cost has collapsed.
The longer take on what changed and why this is possible now →
Draft placeholder portrait
25 years building software.
Now I work for you.
I'm Landon Kuhn, based in Portland, Oregon. For 25 years I've built software — startups, enterprises, and everything between. Production systems used by hundreds of thousands of people. Creative technical work for the craft of it.
Now I work directly with small businesses to build the custom software they need. No agency overhead, no layers of management, no PM playing telephone between you and the code. Just me, your problem, and a fixed price.
The economics of software changed. AI tooling has collapsed what it costs to build custom systems. Work that used to require a team of four and six months, I can often deliver in weeks. I'm passing that on.
The package deal
- → 25 years of engineering judgment. I've seen what breaks and what holds. I build for durability.
- → Modern AI-augmented tooling. What used to take months takes weeks. The economics shifted.
- → Operational accountability after delivery. I keep it running. I don't disappear after the invoice is paid.
The longer take on why each piece is a commodity alone, and what makes the combination at this price new: Why this is possible now →
Draft — placeholder content drawn from prior work
Where the work comes from
The 25 years isn't one continuous track. It's a stack of different work, at different scales, across very different industries — and the patterns transfer downward. Every operational problem a small business has, somebody has already had at much larger scale.
Embedded systems and telecom
Wireless telecommunications devices, real-time fleet tracking, regulatory-compliant logging for long-haul transport, build automation across multi-platform SDKs. The first chapter — where 'production' learned its meaning.
Distributed data, search, and infrastructure
Search and OLAP services, distributed batch pipelines on big-data infrastructure, streaming graph databases, network device management at carrier scale, BI integrations. The unglamorous layer underneath the dashboards everyone reads.
End-to-end product engineering at startups
Robotics telemetry, productivity SaaS, social identity platforms, mobile content delivery and push notifications. A decade of being responsible for the whole system — database, services, frontend, deployment — and operating it through whatever happened next.
AI-augmented platforms and agent orchestration
Retrieval-augmented chat services, multi-agent orchestration, durable agent runtimes in regulated production environments. The most recent layer — and the toolset that makes small-business custom software newly economical.
The instincts come from much larger systems. The shape of the engagement is the only thing that scales down — what holds up under real production load is the same discipline that keeps a small operator's tools boring and durable for years.
How it works
A productized engagement. Defined phases, fixed price, signed scope before I write any code. You know exactly what you're buying before you commit.
01
Discovery
Week 1 — fixed price
A full week understanding your business and the problem. Day 1 on-site if practical. By the end of the week, we sign a written scope — exactly what the build will include, and what it won't. Either side can stop here without further obligation.
02
Build
Weeks 2–3 — fixed price
An 80-hour build envelope, with milestone reviews and one revision pass. Out-of-scope items go through a small change-order process or get deferred to retainer. The price set at the end of Week 1 stays the price.
03
Handover
Week 4
Final delivery into your infrastructure. Documentation, credentials, infrastructure keys — all transferred. The system is literally yours. No vendor lock-in, no withheld access, no dependency on me to keep working.
Discovery is the small, structural answer to skepticism — paid up front, signed scope at the end, either side can walk. The longer take on how the week works →
Keeping it running
After handover, two optional ongoing engagements. Both are standalone products — neither is required to take delivery of the build.
04
Operations retainer
Monthly — optional
Monitoring, maintenance, backups, the occasional small fix. Roughly an hour of attention per month at a deliberately modest price — long-term relationships matter more than maximum extraction from any single line item.
05
Development retainer
Monthly — optional
Prepaid hours for clients who want to keep building. New features, ongoing improvements, iterative investment. Software is never done. The world changes; the software has to change with it.
A modest monthly fee for ongoing access to the engineer who built the system is an unfamiliar shape; the math, the exit, and what the retainer isn’t are all worth being explicit about. The longer take on retainer economics →
What I build
The kinds of systems I build — and where AI earns its place inside them.
Custom web applications
Purpose-built tools designed around how your business actually operates. The music school's scheduler. The publisher's release dashboard. Software shaped to your operation, not the other way around.
Workflow automation
The thing you do by hand every week — reconciling spreadsheets, copying orders into the production schedule, chasing distributors — turned into something that happens on its own.
Data integrations
Connect the tools you already use, or replace three that almost work together with one that fits. Your storefront talks to your fulfillment. Your CRM talks to your billing. Nothing sits in someone's inbox.
Dashboards & reporting
See your business clearly. Real-time data, the metrics that actually matter, updated without anyone asking for a report.
Progressive web apps
Mobile-ready applications that install on any device without an app store. Fast, offline-capable, available everywhere your staff or customers are.
AI as a layer, not a feature
AI is part of how custom software gets built in 2026 — not a chatbot widget, not a separate offering. It's a layer inside the systems above, doing specific jobs where it outperforms hand-coded logic.
A scheduling constraint solver shouldn't be an LLM. A natural-language interface to that solver should be. The judgment is knowing the difference.
Music school scheduler
Natural-language constraint entry — “move the Hendersons to Tuesdays, except the second week of November.” The constraint solver stays algorithmic. The LLM only handles the language.
Publisher's release workflow
Reconciling contributors, attributions, copyrights, and metadata across print, web, and ebook channels. The release state machine stays deterministic. AI handles the messy assembly — names, permissions, version drift.
Cart pod ordering portal
Cross-vendor menu suggestion from informal intent — “something not too spicy with rice.” The order-routing logic stays predictable. AI handles the matching.
Integration pipelines
Structured extraction from PDFs, contracts, scanned invoices. The data flow stays auditable. AI handles the unstructured-to-structured step.
- Wireless telecommunications
- Fleet telematics
- Carrier-scale device management
- Search & analytics services
- Distributed graph databases
- Robotics telemetry
- Social identity platforms
- Mobile content delivery
- Strategic relationship management
- Multi-agent AI orchestration
- Independent music publishing
- Community arts and events
Where this sits
Most small businesses arrive with two reference points for software: cheap-and-generic at one end, and the agency quote that priced you out at the other. For years, the answer has been to pick whichever you least disliked.
Off-the-shelf SaaS
Squarespace, Shopify, QuickBooks, Mailchimp. Fast to start, low price. The cost is fitting your business to the software's assumptions — quiet at first, compounding for years.
→ Where I sit
Custom from a solo engineer
A tier that didn't economically exist a few years ago. One experienced engineer, scoped to your business, built around how you actually operate. The cost-per-unit-of-output dropped enough to make this shape of project real.
Agencies and in-house teams
Right for large-scale projects, deep compliance, or businesses with enough scale to justify a payroll line item. The cost structure — overhead, account managers, layers between you and the work — priced this tier out of reach for small operators in the first place.
The middle didn't used to exist as something a small operator could fund. It does now. The longer take on what fits in this tier →
Is this a fit?
The best projects start with a specific operational problem — something that has been costing you hours every week, something you have been patching with spreadsheets or three SaaS tools that almost work together.
You don't need to arrive with a solution in mind. You need to be able to describe what's broken. I'll figure out whether software can fix it and what that would cost.
If you're not sure whether your situation is a fit, describe it anyway. I'll tell you honestly — including if someone else is the right answer.
This probably isn’t the right fit if…
You need a standard five-page website
A brochure site is a Squarespace project, and Squarespace is genuinely the right tool for the job. I'll say so up front rather than let you pay me to build something worse.
You're looking for a cofounder or startup partner
Venture-backed product startups need a founding team and investor capital. I work with operating businesses that have a specific operational problem — not companies still figuring out what they're building.
You can't name a specific problem
Custom software is built toward something broken. If you can't describe what's costing you time or money, there's nothing to build toward yet. The right starting point is a workflow, not a vague sense that you need "an app."
Heavy regulation at volume
HIPAA, PCI at scale — compliance overhead is real, and a solo shop isn't the right choice when regulatory risk is the primary concern. Some problems need a team with a compliance practice.
Work
Projects that demonstrate the offer.
Agent Orchestration Platform
An agent orchestration platform — built solo, running in production, depended on daily by 20+ operators at a B2B platform in a regulated industry.
Folk Music Publisher Site
A custom publishing and e-commerce site for an independent folk music publisher whose business didn't map onto any existing platform.
Community Booking App
A full-stack progressive web app for a regional arts community — event listings, ticketing, and venue coordination in one system.
Genetic Simulation
A browser-based 3D simulation: agents driven by genetic algorithms, surviving and evolving across generations in a procedurally generated environment.
Questions
How long does a typical project take?
Small projects ship in days to weeks. A full custom application is typically 4–8 weeks from kickoff to launch. I'll give you a real estimate after the initial conversation, and I hold to it. Fixed-price means fixed scope — no scope creep surprises.
What does this cost?
Fixed price, set during discovery. Discovery is a flat fee for Week 1. The build is a flat fee, agreed in writing before any code is written. The operations retainer is a low monthly amount. The development retainer is prepaid hours. No hourly billing on the build, no surprise invoices, no open-ended estimates.
Why is this possible now?
The cost of building custom software has dropped roughly 5–10x in the last two years. AI tooling has changed what one experienced engineer can produce alone. Work that used to require a team of four and a six-month timeline can often be delivered in weeks. I'm passing that on. The economics genuinely shifted — I'm not cutting corners.
How much of this is AI tooling?
Enough to make the economics work. Not enough that it stops being engineering. AI accelerates the typing, not the judgment. I still decide what to build, how to architect it, how it should fail, and how to keep it running. You're hiring 25 years of engineering judgment, applied with modern tools — not generated code with no one accountable for it.
What if you get hit by a bus?
Your business doesn't stop. At handover, the system is literally yours — code, documentation, credentials, infrastructure access. There is no vendor lock-in, no withheld access, no key person risk inside my engagement. Another engineer can pick up where I left off. Keeping me on a retainer is a choice, not a dependency.
What if I need changes after launch?
That's what the retainer is for. Things change. You learn things after launch. A monthly retainer keeps me available without requiring a new project quote every time something needs to change.
What happens if something breaks after you build it?
On a retainer: I fix it. Off retainer: I'll scope it as a small engagement. Either way, you have a real person who built the thing and understands it completely — not a support ticket.
How is this different from hiring a freelancer on a platform?
Accountability. On a platform, you manage the relationship, the requirements, the QA, and the handoff. I handle all of that. I'm a generalist who takes responsibility for the whole system — not just a task. And I stay. Most platform engagements end when the task ends.
Why is this less expensive than an agency?
Agencies charge for overhead: project managers, account managers, sales staff, layers of communication. I have none of that. The price reflects the structure — one person, no margin layer between the work and you, no PM playing telephone between you and the code.
Do you sign NDAs?
Yes, always. Your business information and any proprietary workflow details are confidential.
If you're skeptical
If this sounds too fast or too cheap to be real, I understand.
The economics of custom software changed in the last two years. The shape of the offer is unfamiliar. Skepticism is reasonable. Four places to verify before committing:
- → Why this is possible now The cost shift, what the offer isn't, and what you're actually paying for. The explanation behind the pricing.
- → Recent work Case studies of real shipped projects, in production today. Built solo, kept running.
- → How I work Defined phases, fixed price, signed scope before any code is written. The shape of every engagement is the same.
- → The discovery week A fixed-price first week. Either side can stop after Week 1 with no further obligation. You don't have to trust the claim — only the first small step.
Currently
This site is a work in progress. I'm building it out alongside a few other things and not yet taking on new clients.
If you want to follow along or get in touch when that changes, email is the right place.
landon9720@gmail.com