Peter
Neyra

Product design & engineering
Sofia · Madrid · London
peter.a.neyra@gmail.com

Four products. Several visual systems. One person.

These span two countries, several product categories, and completely different audiences — from Bulgarian food-drop enthusiasts to Spanish house flippers — plus the internal tools I built to run them.


Each has its own design language built from scratch: warm paper and hand-drawn pixel art for a farm-direct food drop; translucent dark glass and WhatsApp-native flows for a renovation marketplace; precision typography and AI-curated data for a house-flipping tool; lifestyle matching and isochrone maps for a property concierge.


Designed and coded end-to-end — no agency, no handoff, no framework I didn't choose deliberately.

01 · Vitrina · 2026

Vitrina
farm drops

Problem. Small farms grow exceptional food but can't reach city shoppers, who pay main-street markups for less. Distribution is the bottleneck.
Solution. Vitrina pools a neighbourhood's demand — people pre-order personalised groceries from small local farms with their neighbours; when an area hits critical mass the drop goes ahead at near-wholesale, collected from a local shop's rented fridge. A phone-first waitlist ranks demand area-by-area before a euro is spent.
Stack. Pure HTML/CSS/JS landing · Supabase · FastAPI on Hetzner · Vercel · bilingual BG/EN. Open it for the live site, the gamified waitlist, and the founder tracker.

HTML · CSS · JS Pixel art No framework Bilingual (BG/EN) Bulgarian market
View project
Vitrina
vitrina.bg
Vitrina — farm drops
Open project

02 · Leo.casa · 2025–2026

Leo.casa
renovation platform

The public platform for Leo.casa — an AI-powered renovation marketplace for Spanish homeowners and tradespeople. Persona-switching landing (homeowners vs contractors), WhatsApp-native contact flows, transparent line-item estimates, verified contractor badges, and a full dispute resolution flow. Bilingual. One coherent design system across every persona and state.

Next.js AI marketplace Homeowners Contractors Bilingual
View project
Leo.casa
leo.casa
Leo.casa
Open project

03 · Leo · 2025

Leo
housing concierge

An AI housing co-pilot for Spain. You don't pick filters — you tell Leo how you live, and it finds the rentals and homes worth your time, with a tailored checklist for every serious viewing. Four personas (renters, buyers, buyer agents, relocation) in one bilingual landing, an isochrone "sweet-spot" map, listing-vs-reality mockups, and a voice-first intake. The most design-dense build of the set.

Next.js Real estate AI Lifestyle matching Isochrone maps Bilingual
View project
Leo · Housing
puedo.casa
Leo — housing concierge
Open project

04 · Leo.casa · 2025–2026

Leo
Estimates

A standalone AI tool for house flippers and investors in Spain. Paste a listing URL, get a room-by-room renovation cost breakdown, comparables of sold properties, an estimated flip margin, and a ready-to-send contractor briefing — in minutes. Designed to feel fast, precise, and trustworthy before the first result even loads. Bilingual (ES/EN).

Next.js AI product House flippers Spain market Bilingual
View project
Leo Estimates
leo.casa/es
Leo Estimates
Open project

05 · Smart Supply Router · 2026

Routing
simulator

Problem. A warehouseless food supplier routes fresh produce straight from farms to restaurants — no depot to buffer demand, so every order is a direct multi-stop run and the margin lives or dies on routing.
Solution. A browser simulator of real farm→restaurant mornings: pick the orders, it pulls real road distances (OSRM), solves the vehicle-routing problem with a self-hosted VROOM engine, and breaks down the true per-drop cost — goods, last-mile, overtime, distance — to show whether the unit economics work before a van leaves.
Stack. Vanilla JS · Leaflet · OSRM road network · self-hosted VROOM solver · scenario data (farms / restaurants / depot).

VROOM OSRM Vehicle routing Unit economics Leaflet
View project
Routing Simulator
VROOM · OSRM
Farm-to-restaurant routing simulator
Open project