Liubov Tsebekova
← All projects

Cine.pick 2023 · rebuilt in 2026

Film discovery shaped by personal taste

Cine.pick began in 2023 as a discovery tool built around niche cinema. During the 2026 rebuild, I opened it to people with different viewing histories and made the resulting taste profile easier to understand.

Product designer & creator · AI-assisted developmentOpen live product ↗Read the outcome ↓
Project
Independent film discovery product
My work
Product model, onboarding, interaction design, visual system and direction of AI-assisted development
Evaluation
Informal usability tests with six participants
Changes
Two entry routes, clearer archetypes and expanded DNA profiles
Cine.pick DNA profile: a recognisable character and a breakdown of film preferences
The updated DNA profile connects a recognisable character to the films someone has rated.
In this case 5 decisions + outcome

01

Make room for different viewing histories

The original starting point assumed people knew niche films. That limited who could begin. I introduced two routes: a familiar selection for a broader audience and a deeper catalogue for specialist viewers.

Both lead into the same rating flow. The choice changes the films someone starts with, while keeping the rest of the experience consistent.

The first choice offers familiar films or a specialist selection.
The first choice offers familiar films or a specialist selection.View full size ↗

02

Ratings distinguish dislike from an unseen film

People needed separate ways to dislike a film and say they had not seen it. I designed distinct like, dislike and unseen actions, and added undo so an accidental rating could be corrected.

I also designed states for no ratings, only dislikes and unseen films. Each explains why a profile is still incomplete and what to do next.

Try a sample of the rating interaction

This example shows the controls and feedback. Your choices here do not update a Cine.pick profile.

Interactive sample · CineMatch controls rate a film
CINEMATCH · INTERACTION SAMPLE

RATE FILMS
ONE BY ONE.

Choose Like or Not for me for films you know. Skip anything you haven’t seen. You can undo the last choice.

Sample
12 films to try
Your choices
Stay in this example
FILMS REVIEWED0 / 12
Possession poster

Possession

Andrzej Żuławski · 1981

NOPE NOT SEEN
Cure poster

Cure

Kiyoshi Kurosawa · 1997

NOPE NOT SEEN
Videodrome poster

Videodrome

David Cronenberg · 1983

NOPE NOT SEEN
Come and See poster

Come and See

Elem Klimov · 1985

NOPE NOT SEEN
Persona poster

Persona

Ingmar Bergman · 1966

NOPE NOT SEEN
Dogtooth poster

Dogtooth

Yorgos Lanthimos · 2009

NOPE NOT SEEN
Under the Skin poster

Under the Skin

Jonathan Glazer · 2013

NOPE NOT SEEN
Stalker poster

Stalker

Andrei Tarkovsky · 1979

NOPE NOT SEEN
Mulholland Drive poster

Mulholland Drive

David Lynch · 2001

NOPE NOT SEEN
The Elephant Man poster

The Elephant Man

David Lynch · 1980

NOPE NOT SEEN
Melancholia poster

Melancholia

Lars von Trier · 2011

NOPE NOT SEEN
Inherent Vice poster

Inherent Vice

Paul Thomas Anderson · 2014

NOPE NOT SEEN
UNDO
NOT FOR ME
HAVEN’T SEEN
LIKE

03

Feedback from six participants informed the taste profile

I ran informal usability tests with six participants. Afterwards, I revised the imagery and copy for the characters and taste archetypes, and added sections to the DNA profile.

The profile gives a short description first, then exposes the themes and taste dimensions behind it. The character provides a recognisable entry point; the breakdown lets someone examine how their ratings are represented.

04

Carry the profile into the next film choice

A taste profile is useful when it helps someone decide what to watch. I added mood-based discovery, more detailed search and connections between films and books.

People can follow a short mood questionnaire or browse through filters, open a film and save it. These routes offer different levels of effort after the initial ratings.

05

Keep the visual system consistent

I carried the black and yellow identity through rating, discovery and film details. A shared icon set pairs symbols with labels, so the meaning of an action does not depend on colour alone.

The same hierarchy and control patterns continue across desktop and mobile.

Outcome

Where the work stands

The rebuild supports familiar and specialist entry routes, separate like, dislike and unseen actions, and a taste profile with clearer character descriptions and a detailed breakdown.

Scope & next steps

Feedback from six participants informed the character imagery, copy and profile structure. Next I want to test the revised profiles and compare the suggestions with what people would actually choose to watch.

Explore Cine.pick ↗

Open to Product / UX / UI roles

Let’s talk about your product.

Based in Lisbon. Looking for a team in Europe, remote or hybrid.

Get in touch