jua.ai

jua.ai builds products on top of weather models. We helped turn a strong technical foundation into the first working web version: an interface where users choose an area on a map, read a forecast, compare signals, and set alerts without relying on internal tools. We worked directly with founder Marvin Gabler. The result was a product that looked and behaved like a real first version, not an internal prototype. It helped jua.ai move toward first customers and supported the story around a funding round of about $2.5M.

~$2.5M round supported by the early version
MVP first usable web version
AWS infrastructure for customer access
Challenge

A strong model needs a clear interface

jua.ai had a serious technical foundation. But users judge the product by how quickly they understand it: which area they are looking at, what the forecast says, where the signals change, and which alerts they can configure.

The job was clarity. The map, charts, states, alerts, access, and infrastructure needed to work as one product.

  • Clear area selection and forecast exploration on a map.
  • Charts and states that make the data easier to understand quickly.
  • Alerts, login, and customer access so the product could move beyond internal use.
Search, map, forecast parameters, and timeline are visible in one workspace.
The early interface already showed the shape of a full product, not a raw model output.
Solution

A web product for forecasts and alerts

We built the first version as one product, not a set of technical pages. Maps, forecasts, charts, and alerts had to feel connected and guide the user without extra explanation. A user could choose an area, read the forecast, understand the movement, and set monitoring in the same interface.

Behind that simplicity was the heavy product work: API integration, auth, customer access, notifications, deployment, and AWS infrastructure.

  1. 01

    Forecast interface

    Map, forecast screens, charts, selected areas, and clear states for everyday use.

  2. 02

    Alerts

    Condition setup, statuses, and monitoring flows that connect forecasts to action.

  3. 03

    Access

    Login, customer keys, limits, and API access for controlled product usage.

  4. 04

    Infrastructure

    AWS, API Gateway, notifications, Docker, GitHub Actions, and a release process for a product that could keep growing.

Technical work

The interface and the system behind it

The hardest part was making a live technical system feel simple from the outside. Forecast APIs, map behavior, chart data, alert logic, authentication, and deployment all had to line up. When one layer works separately, the product quickly becomes heavy.

We helped connect those parts into the first product foundation: web app, forecast integration, alerts, auth, controlled access, AWS deployment, and a clear release process.

01

Web product for forecasts, maps, charts, and alerts.

02

Forecast API integration with interface states and selected areas.

03

Auth, customer keys, usage limits, and controlled access.

04

AWS Lambda, API Gateway, Docker, notifications, and GitHub Actions.

05

QA and release process for fast iteration with the jua.ai team.

Stack
Forecast dashboards the main interface
Mapbox maps and data layers
Alerts and notifications risk statuses for clients
Auth client access
AWS Lambda + API Gateway serverless backend
Docker + GitHub Actions builds and deploys
Alert settings combined the map, risk table, toggles, and location settings in one controlled flow.
The events panel showed how monitoring output returns to the user as a readable history.
Small states mattered too: sent alerts, conditions, dates, and location context had to be legible.
Outcome

What changed

01

The product became clearer

Forecasts, map, charts, and alerts started to feel like one interface instead of separate technical parts.

02

Customer access arrived earlier

Login, API access, alerts, and infrastructure gave jua.ai a more direct path from internal development to real use.

03

The early version supported the next stage

jua.ai moved from technical promise to a product people could see and use. This supported the company on the way to a funding round of about $2.5M.

Related work LiftEd AI Platform · EdTechKaizen Club · TheNext AI Infrastructure · Telegram Mini-App · EventsAI CRM AI CRM · RAG · US market
Services MVP and product launches for founders MVP and Product LaunchesAI for departments: sales, HR, and support AI for DepartmentsAI development for how business works in Kazakhstan AI Development

Turning complex technology into a clear product?

We help teams turn models, data, and infrastructure into a web product where interface, backend, access, and deployment work together.

View
Brief (optional)