LaneBuild
You have an idea and need it built, approved, and live.
Mobile, web, backend, AI, hardware — or all of them. Architecture, spec, build, certification, launch, handover.
Technical architect and hands-on builder for products that have to work in the real world. Apps, AI systems, hardware — from idea to certified and live, and rescue when they fail.
Book a scoping week→Mobile, web, backend, AI, hardware — or all of them. Architecture, spec, build, certification, launch, handover.
Vision, detection, anomaly, LLM integration — with edge hardware, latency, false alarms, and data constraints treated as first-class, not afterthoughts.
Diagnose from evidence. Redesign the part that's actually broken. Ship the fix without a retraining cycle if one isn't needed.
The client couldn't find HR software built for the reality of a workforce deployed into fragile states. Built it end to end over eighteen months: full employee lifecycle for a distributed field workforce — profiles, timesheets, leave, expense and medical claims, performance, payroll data — plus a Safety module that is the point of the app.
A panic alarm that fires offline and delivers the moment connectivity returns. Movement-risk approvals routed through field security officers. Security-alert broadcasts with staff check-in and crisis-response monitoring. Geolocation and incident reporting.
Same discipline as the vision work below, in a different domain: a system that has to be right in the field, under bad conditions, or someone gets hurt.
Built from zero over roughly eighteen months: video and in-person doctor booking, real-time doctor ETA, at-home lab sample collection, e-prescriptions, doorstep pharmacy fulfilment, insurance integration.
Took it through the full health-authority approval and certification process, launched it, and handed it to the client's own team to run. Regulated healthcare approval is a gate most agencies never clear.
Still on the stores, still rated 5 stars regionally, years after handover. That it survived without me is the proof.
A charging pad had to detect foreign objects before and during charging. A missed detection is a safety incident; a false alarm is a session that never starts. It was failing in both directions at once: at night, headlight and torch reflections on the glossy pad fired constant alerts. In daylight, a hand covering a third of the pad scored 0.303 — below threshold — while a small dark bolt scored 0.921.
The team's proposed fix was to lower the threshold. That trades a safety miss for a flood of false alarms. The real causes were architectural: OR-based decision logic with a 7–10px minimum blob, auto-exposure locking after one second, no spatial persistence, per-frame heatmap normalisation, CLAHE washing out the brightness cue, and an anomaly model that had learned large soft bright shapes as normal — because the glossy pad mirrored car undersides and ceiling lights throughout dataset collection.
The fix was a two-tier detector. Tier one is model-free — a log-ratio comparison against a reference frame that catches large objects within two frames regardless of model confidence. Tier two keeps the anomaly model with AND-logic across size, peak, not-glare, and persistence, reserved for small FOD where the model is genuinely strong.
Camera hardware and runtime were identified from screenshot evidence alone — the magenta IR cast, the OpenCV window chrome, and a ~1.5 FPS loop — before touching a line of code. False-alarm detection efficiency raised to 98%. Deployable the same night, no retraining cycle.
Architected the technical approach for a state-scale law-enforcement video platform: automatic number-plate recognition and face recognition across 1,400 cameras, with edge inference per camera rather than centralised GPU processing, event-triggered retention rather than full archival, and integration into national vehicle and identity databases.
Delivered the full cost and architecture model — three-year TCO, per-camera economics, and the build-versus-buy analysis — alongside a live six-camera reference deployment.
Full firmware specification for a three-phase energy monitoring device, including a hardware review that surfaced design issues before they reached production — a reversed isolator channel, an oscillator marking off by a transposed digit, and a boot-strap conflict on a shared SPI pin.
You don't have to trust me with the whole project. Start with a week. For a new build, that week turns your idea into an architecture, a written spec, and a fixed quote. For a system in trouble, I review your code, hardware, and field evidence and tell you what's actually wrong.
Then fixed price, milestone-paid, delivered in a single pass against a written definition of done. There is a method behind that — I ran a delivery framework across nine years of agency work — but the method isn't the product. The shipped thing is.
I diagnose from evidence before touching code. I use AI-assisted tooling heavily to compress build time; what you're paying for is judgment and architecture, not typing hours. And there are no free consultations — the scoping week is paid, or it doesn't happen.
Architecture, AI/ML, deployment. A senior technical owner on call for the team that already has hands but not the judgment layer.
Fixed price, milestone-paid, single pass to a written definition of done. Scoped from the week below.
One week, fixed fee. New build: architecture, written spec, fixed quote. System in trouble: code, hardware, and field evidence reviewed — and what's actually wrong, in writing.
Tell me what you're building or what's broken. I'll reply with whether the scoping week is the right first step, and if it is, when it can start.
ankit@utenx.com