Reliable code. Real impact.
Penumbra Tech is the single-engineer practice of Wesley Weaver Jr. I build across the seams: software, infrastructure, data, operations, and the human process around them. Web apps, AWS, Unreal Engine systems, and the reliability work that keeps everything from quietly falling over.
The actual code I write.
This is a real function from the diagnostics dashboard: the SSE flush that lets the latency chart stream cleanly even when k6 is firing 1,500 req/s. Tight, documented, no framework gymnastics. That's the style across every project on this site.
Three ways to work together
Concrete shapes with concrete prices. Pick one or ask me which fits — I'll tell you if neither does.
Diagnostic Week
A short, focused engagement to find your actual constraint and put a written plan against it. One week digging through your stack, your data, and the bottleneck you can name. End the week with a written plan plus a fixed-price build quote. The fee rolls into the build credit if we move forward; the plan is yours either way.
Best for: You think you know what is wrong, but you want a senior eye on it before committing to a 4-6 week sprint.
Build Sprint
4-6 weeks of fixed-scope work with weekly demos. Half paid up front, half on acceptance. The second half is not owed if week-one deliverables miss the agreed criteria. Typical scopes: a customer portal, an internal dashboard, a backend rescue, an audit and remediation.
Best for: You have a defined project and you want it shipped reliably without having to manage it daily.
Retainer
Block-of-hours arrangement for ongoing operations, on-call backstop, smaller asks, or the long-tail mode after a build engagement wraps. Hours roll month to month; no minimum monthly commitment.
Best for: You want a senior engineer on speed-dial without hiring full-time.
Next intake window opens within ~2 weeks of intro call
"Heyo! Ryan from 360 Auto here. Our website was in shambles and Wes was able to redesign it and brought up the calls and car count from 20-30 up to 40-50 a week. He treated our website like his own and really knocked it out of the park."
Ryan, owner · 360 Auto · Port Orchard, WA
Read the full engineering writeup: Repair360 case study →
Engineering across the stack
Most projects need a generalist who can pick up the entire problem instead of handing it off across three teams.
Custom software
Full-stack web apps, internal tools, and APIs. Built clean, documented, and handed over so your team can keep going.
Cloud & backend
AWS deployments, MySQL and DynamoDB, CI/CD pipelines, and the operational glue that keeps a service running at 3am without paging anyone.
Game development
Unreal Engine builds, multiplayer systems, AI behaviour, web portals for live games. See the case study under Projects.
Performance & reliability
Load testing, security hardening, incident response. I find where your stack breaks before your users do.
Evidence, not promises
Each project below is live and exercisable. Click through to see the real working software.
Pick whichever question you arrived with.
Most prospects arrive with one of three unasked questions: what does this cost, why pay for an engineer when an LLM exists, or could I just do this myself. There's a long-form answer for each, so the 30-min intro can be about your project instead.
Have something tangled you want untangled?
Send a quick note describing what you're building or what's stuck. I'll reply with whether I think I can help. No commitment, no template-bot follow-up sequence.