One person. On purpose.
LanceTek isn't a studio pretending to be bigger than it is. It's one software engineer, working directly with the owners and staff who'll use what gets built.
The person you speak to on the first call is the person who writes the code, sets up the server and answers the phone when something breaks.
That's the whole point. There's no account manager translating your problem for a junior developer, and no handover between sales and delivery where the detail gets lost.
The trade-off is honest: I take on a small number of projects at a time, and I'll tell you if I can't start soon. In return you get continuity, someone who remembers why a decision was made, and prices that don't carry an office and a sales team.
- Boring technology, on purpose.Well-known languages and databases with long support lives. The clever part should be the fit to your business, not the stack.
- Written down, always.Scopes, decisions and changes go in writing. You should never need to remember what we said on a call.
- Small first, then bigger.A paid discovery or pilot before a large build, so you can judge the work before betting on it.
- People approve, machines draft.Automation does the reading and the typing. Decisions that matter stay with someone accountable.
- Easy to leave.Your accounts, your code, your documentation. If you move to another supplier, it should be a short email, not a rescue.
What things are built with.
Chosen for long support, good documentation and a large pool of other developers who know them. Each one has a Journal entry explaining why, in client terms.
Build
- TypeScriptweb apps and APIs
- Pythondata jobs, document handling
- PostgreSQLthe database for nearly everything
- SvelteKitinterfaces staff actually like
Automate
- n8nworkflow runner, self-hosted
- Hosted LLM APIsfor drafting and extraction
- Local modelswhen data shouldn't leave
- Approval queuesbuilt per client
Run
- Linux and Dockereverything containerised
- Proxmoxvirtual machines on owned hardware
- Caddyweb server, automatic HTTPS
- WireGuardprivate remote access
Watch
- Uptime checksalerts to a phone, not a dashboard
- Grafanagraphs for the curious
- resticencrypted, tested backups
- Gitevery change recorded
Tested at home before it reaches you.
A small rack in the office runs the same tools I recommend to clients. Updates, backups and failures happen here first, so the surprises are mine rather than yours.
- 01Firewall and router. Segmented networks: clients' test systems never share a network with anything else.
- 02Three small compute nodes. A cluster for virtual machines; one can fail without anything stopping.
- 03Storage with mirrored disks. Snapshots every hour, kept for a month.
- 04Off-site, encrypted backup. Restored and checked monthly, the same routine clients get.
Ask me anything before you commit.
Including the awkward questions: what happens if I'm ill, how I price, and what I won't take on. The first call is free and there's no pitch.
Start a conversation