My Role

Key Contributions 

• UX Research & Design 

  • Conducted user interviews with 5 startup employees to map their current video workflow.

  • Designed a frictionless, single-step flow - from text input -> AI generation -> video preview.

  • Focused on simplicity: clean layouts, minimal input fields, and strong visual hierarchy for clarity under time pressure.

I contributed as a UX/UI designer and front-end developer, bridging user experience and technical execution to make AI video generation feel effortless.

• Front-End Developer

  • Implemented the React-based interface with responsive components for upload, preview, and playback.

  • Integrated Gemini API + OpenAI API calls to display real-time AI outputs (script + visuals + voiceover)

  • Collaborated with backend engineers to connect the UI with AWS services (S3, Lambda) for instant rendering

Startups relay on short-form videos to reach their audience - TikTok ads, Instagram Reels, YouTube Shorts

But producing those videos consistently is a major challenge

"We need new videos every week, but we don't have the time or budget to make them."

1. Too Expensive - Even a 1-minute product video can cost hundreds to thousands of dollars when outsourced.

2. Too Slow - 2-3 weeks per production cycle

3. Too Complex: Each platform needs different formats, captions, and aspect ratios

Glöd AI set out to change that — by letting teams create high-quality, platform-ready videos in minutes, not days.

The Solution

Glöd AI uses generative AI to automate the entire video pipeline:

• Input: Upload product photos and short descriptions.

• AI Engine: Writes the script, generates voiceover, and visual sequence using Gemini + OpenAI APIs.

• Output: A 1-minute, ready-to-post marketing video stored and delivered via AWS.

No editing software. No video expertise required.
Just an intuitive web app that lets startups tell their story — instantly.

Impact & Leanings

What started as a hackathon project became a real tool with real potential for early-stage teams.
Glöd AI reduces production time from 2–3 weeks → under 1 hour, and saves thousands in agency costs.

Through this project, I learned that:

• Good UX is invisible: when automation feels natural, users trust the system.

• Design and code amplify each other: being able to implement my own interface accelerated collaboration.

• AI design isn’t about replacing creativity — it’s about removing barriers to it.

Reflection

Glöd AI taught me how to design not just for users, but for systems that serve them.
By combining thoughtful UX and scalable frontend architecture, I helped turn an abstract AI concept into a tool that genuinely empowers startups to market faster and smarter.

Refining grocery scanning through iteration and user feedback

Through several design iterations, we refined TrackFood’s core experience around effortless grocery input. Our early prototypes explored different interaction models — from manual entry to OCR-based receipt scanning — each revealing friction points that informed the next version. The final hybrid approach, combining auto-recognition with simple manual edits, achieved a 90% success rate during usability testing. Users described it as “intuitive and natural, like taking a quick photo of my shopping list.”

Inventory Dashboard & Expiration Alerts

The dashboard organizes items into Expiring Soon, Recently Added, and History—highlighting urgency instead of location. Subtle color cues and reminders guide users to use items in time, turning awareness into simple, everyday action.

Smart Scan

Smart Scan lets users capture grocery items instantly using receipt or barcode scanning. With OCR technology, the app reads item names and predicts expiration dates automatically. Users can review or edit details before saving, making the process fast, accurate, and effortless.

AI Recipe Generator

The AI Recipe Generator suggests meals using ingredients already on hand. With one tap, users can generate custom recipes that reduce waste and make cooking more creative and sustainable.