AI Designer & UX Engineer in Dubai
Haroon Al Hadi (Hadi) designs intuitive interfaces for autonomous agents, generative models, and non-deterministic software—transforming complex LLM capabilities into seamless, high-confidence user experiences in Dubai and worldwide.
Trusted by high-growth startups and global product teams:
Proven Systems & Thought Leadership
Selected Work & AI Architectural Frameworks
Case studies and published architectural playbooks on multi-agent UIs, non-deterministic states, and generative interfaces.
40% Faster AI Delivery · +19% Task CompletionEducore Systems (Dubai & Global) · Enterprise AI & Workflow Systems
AI-Accelerated Platform & Intelligent Matching
Designed intelligent persona matching algorithms, automated applicant scoring interfaces, and AI-accelerated administrative workflows across 15+ educational modules.
View Case Study
+28% Conversion · +38% Mobile POS UpsellLA Fitness (USA & Canada) · Intelligent SaaS & Funnel Optimization
Predictive Lead Scoring & Sales Automation
Architected contextual recommendation engines and predictive lead scoring flows that enabled sales agents to close memberships with intelligent real-time guidance.
View Case Study
Complete 0-to-1 Mobile & Flutter BuildIndependent iOS Product · Tactile Mobile & Geofencing UX
NapWake Smart Spatial Alarm
Designed an intelligent, context-aware mobile experience with fluid micro-interactions, haptic feedback, and background state management.
View Case Study
+23% Online Sales · -31% Checkout FrictionWhite Mammal & Impresa · Conversion Funnel & Dynamic UX
Algorithmic Discovery & Checkout Optimization
Engineered high-velocity personalized product recommendations and frictionless one-tap checkout flows from social discovery to purchase.
View Case StudyPublished AI Design Frameworks:
Designing for Autonomous Agentic Workflows & Multi-Agent UIs
How to design supervisor loops, state machine transitions, and reversible action gates that reduce cognitive load in multi-agent software.
Read Article Human-AI InteractionDesigning for Non-Deterministic Interfaces: Latency & Hallucinations
Visual hierarchy mapped to model certainty thresholds, progressive cognitive breadcrumbs, and deterministic fallback guardrails.
Read Article Generative UIFrom Static Interfaces to Generative UI: A Product Designer's Playbook
Architectural patterns for streaming dynamic components and just-in-time interface generation from structured LLM tool outputs.
Read ArticleRecommendations
What Product Leaders & Engineers Say
Verified testimonials from product designers, developers, and engineering collaborators.
“Hadi is a highly insightful individual with a natural ability to identify and empathize with user pain points. What truly sets him apart is his inherent curiosity and drive to not only understand these challenges but to actively seek and test creative, out-of-the-box solutions. His enthusiasm for learning and his proactive approach make him a valuable contributor to any team focused on user experience.”
Verified on LinkedIn“Hadi is the person who introduced me to Product design. He is a person with deep and specific knowledge in all aspects of design. He creates intuitive solutions for all the hurdles and I have had the opportunity to see that first hand and learn from him. He also has a good ability to articulate his workflow and knowledge to the rest of the team and stakeholders! I recommend Hadi with full confidence.”
Verified on LinkedIn“Hadi is an exceptional UX designer whose technical expertise and clear communication have significantly enhanced our team's success. His ability to transform complex challenges into user-friendly solutions is truly remarkable.”
Verified on LinkedIn“A talented UX Designer who combines creativity with technical expertise to deliver outstanding user-centric solutions. Haroon's collaborative spirit and dedication to excellence make him an invaluable team member.”
Verified on LinkedInAI Design Capabilities
Specialized AI & Agentic UX Services
Tailored capabilities to help your team design, test, and ship modern AI products that users trust.
Agentic Workflow & Multi-Agent UI Architecture
Move beyond linear prompt-response loops. I design supervisor workspaces, agent handoff visualization, sub-task parallelism, and clear progress states that give users full visibility into complex autonomous operations.
Key Deliverables:
- Multi-Agent Supervisor Workspaces
- Agent State & Parallel Task Visualizers
- Human-in-the-Loop Approval Gates
- Reversible Action & Rollback Cards
Generative UI & Dynamic Component Streaming
Transform plain text LLM responses into rich, interactive components. I architect structured JSON tool calls that stream live editable charts, interactive tables, dynamic forms, and transaction cards on the fly.
Key Deliverables:
- Tool-to-Component Mapping Specs
- Streaming Skeleton & Partial Render States
- Contextual Inline Micro-Inputs
- Dynamic Dashboard Layout Engines
Non-Deterministic UX & Confidence Hierarchy
Design UI patterns that respect probabilistic uncertainty. We map visual contrast and action thresholds to model confidence levels (>85%, 50–85%, <50%), preventing costly hallucinations and building user trust.
Key Deliverables:
- Confidence Score Badges & Color Tiers
- Source Citation & Grounding Overlays
- Ambiguity Disambiguation Prompts
- Deterministic Fallback Controls
Perceptual Latency & Cognitive Feedback
Solve the 5–20 second agent latency bottleneck. I design active cognitive breadcrumbs, live thought-stream summaries, and speculative UI rendering that keep users engaged during heavy model execution.
Key Deliverables:
- Reasoning Chain Summaries (Collapsible)
- Live Tool Execution Indicators
- Optimistic UI & Cache Pre-fetching
- Graceful Timeout & Retry Patterns
AI Design Systems & Token Architecture
Build scalable Figma and code component libraries purpose-built for AI products. Includes multi-modal audio/vision states, streaming indicators, token counters, and prompt input affordances.
Key Deliverables:
- AI Interaction Figma Tokens
- Dark & Light AI Component Kits
- Voice & Multi-Modal Waveform States
- Tailwind / React / Flutter Tokens
Fullstack AI Frontend Prototyping (Next.js & Flutter)
Bridge the gap between design and engineering. I build functional, high-fidelity prototypes wired directly to LLM SDKs, allowing stakeholders and beta users to test real latency and workflows immediately.
Key Deliverables:
- Working Next.js / React Web Prototypes
- Flutter iOS / Android AI Prototypes
- OpenAI / Anthropic / Gemini SDK Wiring
- Production-Ready Component Handoff
Methodology
How We Design AI Products in 4 Sprints
A structured, iterative workflow designed specifically for the unique challenges of probabilistic models.
Model & Capability Discovery
Audit target LLMs/agents, context limits, latency benchmarks, tool-calling APIs, and failure modes to define the interaction boundaries.
Agent Flow & State Architecture
Map agent loops, supervisor delegation, human-in-the-loop checkpoints, and low-fidelity state machines to validate user control.
Generative UI & Design System
Craft high-fidelity generative UI components, streaming animations, confidence tiers, and comprehensive Figma design system tokens.
Interactive Code Prototype & Handoff
Deliver production Figma files and working Next.js/Flutter code prototypes with full developer documentation and prompt specs.
Engagement Models
Flexible Ways to Collaborate
Select the engagement model that fits your product stage and delivery timeline.
AI 0-to-1 MVP Sprint
Timeline: 3 – 4 Weeks
Complete end-to-end design for a new AI application, autonomous agent interface, or generative workflow ready for engineering build.
- ✓ Complete agent workflow & state architecture
- ✓ Generative UI card system & component kit
- ✓ High-fidelity Figma specs for web and mobile
- ✓ Interactive prototype with streaming animations
- ✓ Developer handoff & prompt interface specs
Fractional AI Design Lead
Timeline: Monthly Retainer
Embedded senior AI product designer working directly with your founders, AI researchers, and engineers week over week.
- ✓ Dedicated weekly AI design & architecture bandwidth
- ✓ Direct integration into Slack, Linear, and GitHub
- ✓ Continuous feature design & agent loop iteration
- ✓ Design system governance & evaluator feedback
- ✓ Bi-weekly strategic syncs & roadmap alignment
AI UX & Latency Teardown
Timeline: 1 Week Delivery
A rapid diagnostic of your current AI tool to fix user drop-off, perceptual latency, confusing error states, and low trust.
- ✓ Audit of prompt UX, latency states & agent loops
- ✓ Heuristic evaluation of hallucination risk
- ✓ Prioritized wireframes for streaming & confidence UI
- ✓ Quick-win Figma component fixes
- ✓ 60-minute video walkthrough & implementation guide
FAQ
Frequently Asked Questions About AI Design
Common questions on non-deterministic UX, multi-agent workflows, code handoff, and collaboration.
Traditional UX is deterministic: a button click always yields the exact same state. AI interfaces are non-deterministic and probabilistic. AI product design requires architecting for latency, confidence scoring thresholds, dynamic generative component streaming, context window boundaries, and human-in-the-loop (HITL) error recovery loops.
Let's Build Your AI Interface
Whether you are architecting a multi-agent orchestration tool, a vertical AI copilot, or generative UI workflows, let's turn your model into a product users love.




