ROLE
Sole Product Designer
TOOLS
Figma, Webflow, Midjourney, Gemini
SKILLS
User Research, Wireframing, Prototyping, Usability Testing
(01)
Challenge
OORT is an AI-focused blockchain company building a platform to collect high-quality datasets for training AI.
The vision was to leverage community power: users complete micro-tasks, contribute valuable data, and get rewarded — starting with data collection.
(02.01)
Discovery
Competitive Analysis
Existing platforms like Clickworker or Toloka offer small rewards for simple tasks, but being Web2-centric, they result in less diverse datasets. DataHub aimed to solve this with a Web3-native approach.
During discovery, I worked closely with stakeholders, PMs, and engineers to align on the vision and understand the technical boundaries of blockchain. I also conducted workshops and research to map out the opportunity space.
Key insights included:
Target customers: AI startups.
Contributors: Early adopters were Web3 users in developing countries, with later expansion to Web2 users as the product matured.
Data contribution focus: High demand + lower learning curve for contributors.
Blockchain integration: Every data point had to be stored on-chain, ensuring transparency and trust.
(02.02)
Ideation
The first challenge was integration: OORT already had its own wallet app — Ale Wallet. The idea was to embed DataHub within it, but DataHub was also designed to function as a standalone product with its own flows and settings.
At this stage, the priority was validating hypotheses — not polishing the UI. The key question was: how can DataHub live inside the wallet while staying aligned with blockchain mechanics?
For the Alpha version, I decided to keep Ale Wallet’s design and integrate DataHub as a separate navigation item with its own flow. This allowed us to test the concept quickly without fragmenting the experience.
(03)
Datahub experience
At its core, DataHub is about making data contribution effortless. As our community grew, the number and variety of tasks we could launch expanded as well.
My focus across iterations was to reduce friction for contributors while ensuring the quality of the datasets.
To achieve this, I refined the home screen to highlight what users care about most: their progress and available tasks. I introduced streamlined cards, intuitive filters, and clear step-by-step instructions, making the experience faster, simpler, and more rewarding.
Rewards motivate users — but competition motivates them even more. Traditional platforms rely on flat reward systems, which quickly become repetitive and uninspiring.
To keep users engaged, I designed a competitive reward mechanism where contributors earn points for completing tasks, daily logins, and other activities. These points fuel leaderboards and monthly competitions, giving users a reason to come back and push further.
I also introduced a referral system to make the competition fairer and more rewarding. This not only increased satisfaction but also helped grow the community organically.
One major upgrade was the introduction of Launchpad — a feature that turned DataHub into a platform where anyone could launch dataset collection tasks.
Working closely with the PM and developers, I designed Launchpad to make it simple and intuitive for clients to create tasks and publish them directly from the DataHub app.
This gave organizations a self-serve tool for launching their own data collection campaigns, dramatically expanding the platform’s flexibility and scalability.
(04)
quality assurance
Crowdsourced data always carries risks: users can upload incorrect images, create duplicate accounts, or even submit AI-generated content.
To address this, I collaborated with a dedicated QA team and designed a platform that boosted their efficiency 10x, enabling faster and more reliable data-quality control.
This ensured that DataHub delivered not just large datasets, but trustworthy datasets — a key requirement for training AI.
(05)
Visual Design
Through multiple iterations, I created a unique design language that resonated with OORT’s brand while still feeling approachable and enjoyable for everyday users.
I developed a scalable design system in Figma, built on a clear structure with diverse, reusable components to support fast iteration and consistency.
The final visual direction strikes a balance between a minimalistic AI aesthetic and the playful, community-driven energy of Web3, giving DataHub both credibility and personality.
(06)
Results
Scaling DataHub into the leading decentralized data-collection platform
After nearly 18 months of continuous iteration and refinement, OORT DataHub established itself as a category leader in decentralized data collection.
Users registered since launch
Average weekly active users
Average rating of the app in AppStore