Fullstack Dashboard MVP

I built the first web-facing MVP for Altaroad's nanomat concept—Django end-to-end so pilot users could see ML-derived truck weights and basic site context as vehicles exited construction sites, not only raw pipeline logs.

I focused on a thin but real vertical slice: persistence and views that could evolve into the later Vue production suite, with clear enough ergonomics for demos to clients and internal field teams.

Altaroad was proving an ML-powered weigh mat ("nanomat") to estimate truck weight at construction exits; stakeholders needed a credible browser experience to validate the idea alongside hardware and models.

Without a dashboard, weight readouts lived in scripts or ad-hoc exports, which slowed pilots and made it harder to compare sites or time windows.

Design and ship a fullstack MVP dashboard in Django and Python: sensible data modeling, server-rendered or app-style pages as appropriate, and enough structure for the team to iterate toward production UIs.

Implemented Django models, views, and templates (or early APIs where needed) to surface weigh events, vehicles or tickets, and site metadata from the pilot integration path.

Prioritized read-heavy operator flows, empty and error states that made bad ML days legible, and patterns the 2019 Vue dashboards could later reuse without a ground-up rewrite.

The MVP made nanomat outputs demoable and operationally visible during early field pilots, de-risking customer conversations before heavier frontend and multilingual work landed.

Engineering gained a concrete vertical slice from field signal to screen, which shortened feedback loops between ML tuning and what operators actually saw at the gate.