Kulkita
What it is
A web app for stock and freshness in a kitchen, built for the kitchens that cook Indonesia's free school meals.
Who it's for
Kitchen staff who log stock every day, and the head of the kitchen, who needs to know what is in stock and what will spoil soon.
How it works
Each batch is logged in and out with its date, the app flags what is close to spoiling, and it forecasts the next order.
The problem
Context
The free nutritious-meal programme (MBG) runs through local kitchens (SPPG) that handle large volumes of fresh food. Manual or scattered stock records lead to spoilage, wrong stock counts and food-safety risk.
What we learned first
The first thing a kitchen manager needs is to know what they have and what is about to go bad. Prediction only matters once that view is trusted.
Constraints
- Hackathon deadline
- Users are kitchen and logistics staff who rarely use apps at work
My role
Richard on this project
Founder, Product Manager & UI/UX
Product skills
- Product vision
- PRD writing
- Design thinking
- UI design
- MVP scoping
Project skills
- Scrum leadership
- Pilot & impact measurement
- Pitching
As product manager
- Wrote the product vision and the PRD the team built from
- Ran design thinking, then turned it into the UI
- Cut the MVP to visibility first, prediction later
As project manager
- Led a 4-engineer Scrum team from hackathon to pilot
- Ran the pilot at UB Coffee and measured waste for two weeks
- Owned the pitch deck and the final pitch
The work
Key decisions & trade-offs
- Decision 01
Visibility first, prediction second
We cut the MVP to real-time stock visibility and freshness tracking first, then layered spoilage and demand prediction on top.
Trade-off A less flashy AI story in the demo, but a product a kitchen could start using on day one.
How it was built
Stock-in and stock-out logging, freshness status per batch, spoilage alerts and demand forecasting. After the hackathon I kept building it as a founder and ran a pilot in a working kitchen, UB Coffee.
- Stock inBatch & date logged
- FreshnessStatus per batch
- Spoilage alertUse this first
- ForecastWhat to order next
- −35% waste2-week pilot
From the work itself
The result
- 35%
- less food waste in the first 2 weeks of a pilot at UB Coffee
- IDR 26M
- non-dilutive funding from the two awards
- 1st + Top 3
- ElevAIte Brawijaya Hub and Best AI Ideation (Microsoft × Yayasan BUMN)
What I'd do differently
Winning taught me to pitch empathy. Judges remembered the kitchen far more than the model.