top of page
High fidelity mockups
Client App: Project Configuration
The client app lets PMs and data scientists configure, launch, and monitor annotation projects entirely on their own. The core design challenge was making complex workflows, multiple annotation stages, custom ontologies, tiered QA pipelines approachable without oversimplifying them.
I designed a structured project setup wizard that walks clients through defining tasks, building their label taxonomy, configuring multi-stage workflows, and setting a budget, all before any annotation work begins.
A credit-based billing system and pre-launch budget estimator gave clients clear cost visibility, with all payment friction consolidated into a single launch moment rather than scattered across the setup flow.
Design Process
Working in a fast-moving startup environment, the engagement demanded a lean, iterative design process:
Structured requirement workshops to align on scope and priorities
Rapid design cycles where prototypes and UI builds replaced lengthy spec documents.
Requirements evolved through the work itself
each round of screens surfaced new edge cases and product decisions, which were resolved in real time rather than upfront.
Talent app - Annotation flow
The talent app is where annotators do the actual labelling work, and it had to work equally well for expert ML practitioners and novice crowd workers.
I designed a focused, task-by-task interface that surfaces only what annotators need in the moment — the asset, the tools, and the label set — with strong progress indicators and auto-save to support long annotation sessions.
Relationship annotation (linking two spans or objects) was redesigned from a keyboard shortcut into a visual mode any skill level could discover and use. A qualification testing flow, with clear pass/fail feedback, gates annotators into projects automatically — no manual client review required.
High fidelity mockups
Collaboration
General Manager
Engineer Team
Lead Product Designer
DURATION
12 months, Full-time
ROLE
Lead UX Designer
PROJECT
AI Labelling app
CLIENT
Toptal, AI Engeneer Team
Project detail
Toptal needed a production-grade annotation platform that would let enterprise clients build high-quality training datasets for AI/ML models, without requiring engineers to configure every workflow. The platform had to serve two distinct user types with fundamentally different needs: clients (PMs, data scientists, ML engineers) who configure and manage projects, and annotators (expert crowd workers with varying technical levels) who do the actual labelling work.
I was brought in as the lead UX designer to design the full product experience: from initial project creation and workflow configuration through annotation task interfaces, qualification testing, QA review, billing, and results export.
AI Labelling & Annotation Platform
bottom of page