AI Task Management & Scheduling for Professional Kitchens
Yordando is an AI task-management and scheduling system for restaurant kitchens.
Line cooks carry a small handheld e-paper device. Every time a task is finished, they press one physical button. Those presses become the live input for a central AI engine that continuously re-optimises who does what, and when — and delivers the result to head chefs through a web dashboard.
I joined as the founding designer and stayed through to market entry: I co-invented the product (PCT international patent WO2025126247), designed both of its digital interfaces, ran the research with the chefs who would use it, and led the design work that took it to investors, trade fairs and the first 100 production prototypes.
Lead with Product Designer. The executive work is a strength, but it should read as “a designer who could carry the product all the way to market”, not as a second, competing job title. Section 8 handles that.
Professional kitchens still run on memory and shouted instructions. Shifts are planned from experience, tasks are distributed verbally, and during peak service the plan and the reality drift apart within minutes: work stacks up on one station while another idles, and new cooks lose track of what comes next.
Scheduling software exists, but it lives on the manager's laptop. It is structurally disconnected from the person standing at the pass with a pan in each hand. It can record a plan; it cannot see a service.
Yordando closes that loop by moving the feedback point down to the line cook — a handheld e-paper device with physical buttons. One press per completed task. The data flows back to a central AI that adjusts distribution and scheduling in real time, and surfaces the reasoning to management through a SaaS dashboard.
My responsibility: the two digital interfaces — the constrained on-device UI and the SaaS web application — together with the physical form of the device and the field research with end users.
Tools: Figma (UX/UI, prototyping, design system), semi-structured interviews and in-kitchen observation.
The design problem was not “make a nicer screen.” It was:
How does a cook working in heat, noise and constant motion — hands gloved, wet or full — report progress with almost zero friction, so that the AI schedules from live reality instead of after-the-fact estimates?
This split into two tensions that shaped every decision:
No cook is going to unlock a phone and tap an app mid-service. Hands are dirty, screens are unreadable under kitchen lighting and heat, and any interaction longer than about a second simply will not happen. The input had to survive the environment, not ask the environment to accommodate it.
A system that assigns tasks and records who completed them is, from the cook's side of the counter, indistinguishable from surveillance unless it is designed otherwise. If cooks feel monitored, they resist, or they press the button carelessly. And the moment the feedback data stops being honest, the AI schedule collapses. Trust was not an ethical footnote on this project — it was the load-bearing requirement.
I ran the user research directly, in the kitchens and with the people who would live with the product.
Select the 2–3 that genuinely did — write each as: what we heard → what it changed.
Two users, two interfaces, one loop.
Each decision paired with its reasoning — this is the substance of the case study. Show the artefact next to the argument.
In an environment of grease, gloves, steam and heat, a mechanical button is the only reliable zero-attention input. E-paper stays legible under harsh kitchen lighting, doesn't glare, and draws so little power that the device runs a full service and beyond without charging.
Cooks have no reading time. Information density was pushed to the absolute floor. The refresh characteristics of e-paper enforced this discipline rather than fighting it — the medium made restraint the only workable option.
I designed the web application so that a head chef can read the generated distribution at a glance and adjust it by hand — authority stays with the person who knows the kitchen.
Because I was responsible for the device's physical form as well as its interface, refresh rate, contrast, button count and grip position were decided together with screen hierarchy rather than handed over the wall.
The system: a handheld e-paper device with physical confirmation buttons and a single-task screen, a central AI scheduling engine, and a SaaS web application where head chefs read, adjust and act on the generated plan.
The loop: every button press is a timestamped fact about the kitchen. The engine schedules from those facts. The dashboard explains the result. The head chef corrects it where the model is wrong. The correction becomes data too.
My contribution, stated plainly: I designed a slow, colourless, low-resolution screen into the least intrusive and most trusted feedback point in a high-pressure kitchen — and designed the management interface that made its output legible and overridable on the other end.
Because Yordando was an early-stage company, my role extended past the product itself. I include this not as a separate job, but because it is the reason the design ever met a real kitchen.
I worked with the design and production team in Shanghai (Between Design) to take the device from design files to 100 market-ready 3D-printed prototypes, resolving form, tolerance and finish questions in production.
I produced the visual component of the patent application — technical sketches, rendered figures and the illustrated explanation of how the system works — filed as PCT international patent WO2025126247, on which I am named co-inventor.
I designed the pitch materials and pitched to VC funds in Vilnius and Shanghai, supported the company's EU Next Generation fund application, and arranged founder–investor meetings. I also designed the Y Combinator application deck.
I designed the company's brand and communication system — website, brochures, business cards, presentation templates — and the full exhibition identity (stands, posters, print) for SIAL Shanghai, the Shanghai World AI Conference, Collision Toronto and the China Hi-Tech Fair, Shenzhen. I also produced and edited the video content explaining the solution for these events and for social media.
I led the on-site training of professional chefs using the system — closing the loop from the research that started the project to the people finally holding the device.
In extreme environments, constraints are not obstacles to design around — they are the design direction. E-paper's slowness gave me the discipline I needed. And more fundamentally: whether an AI system works is not a technical question. It is a question of whether the people feeding it data trust what it does with them.
Yordando is why I now want to design human-centred AI systems for food and hospitality — systems where efficiency and the experience of the people inside them are not in competition. It is the same question I pursued in my master's research on the relationship between people and AI, met in a hot kitchen instead of on a page.
Reference blocks and production notes. Turn the appendix off before publishing.
A handheld e-paper device and AI scheduling platform for restaurant kitchens. Line cooks confirm each finished task with one button press; the system turns that live signal into optimised staffing and task distribution. Co-invented, designed end-to-end, and taken to market across four countries.
A one-button kitchen device that feeds real-time task data to an AI scheduling engine. PCT patent WO2025126247.
Product Designer & Founding Team Member · 2022–2025 · Yordando (Vilnius / Shanghai)
Device interaction & form · SaaS web application (UX/UI, Figma) · Field research with 10 professional chefs · Patent visuals · Brand, pitch and exhibition design · Manufacturing liaison for 100 market-ready prototypes