Crafting AI experiences
people actually feel.

Generative AI photo editing didn't exist on any smartphone. The challenge wasn't just building a feature — it was defining what this category should feel like on a phone. How do you make a complex AI capability feel effortless in a tool people use every day?

The stakes were high: this would be Samsung's flagship AI differentiator for the Galaxy Z Flip6 and Fold7 launch — and the first of its kind on any phone, anywhere.

Lead Product Designer — owned the end-to-end UX for Photo Assist, from early concept through to production. Worked directly with the Seoul-based tech team to understand model capabilities and constraints, then translated those into a user experience that felt intuitive rather than technical.

The design process centred on a core tension: the AI could do a lot, but users shouldn't have to know that. I designed the full interaction arc — how users discover the feature within the Gallery, how they apply an edit, how they preview the AI-generated result alongside the original, and how they save or iterate.

Close collaboration with the Seoul tech team was critical. The model had constraints — certain types of edits worked reliably, others didn't. Rather than exposing every capability, I worked with engineering to curate the experience, surfacing what the AI did well and gracefully handling edge cases.

Discovery within the Gallery — Photo Assist lives where photos already live. Rather than asking users to open a separate editing app, the feature is surfaced contextually when browsing photos, reducing friction to near zero.

Preview-first interaction — Users see the AI-generated edit alongside the original before committing. This builds trust in the AI and gives users a sense of control over an otherwise opaque process.

Constraint-led curation — Instead of exposing the full range of the generative model, the UX presents a curated set of edits matched to the specific photo. This kept the experience reliable and prevented the "uncanny valley" moments that erode user trust in AI.

First of its kind — launched as the first commercial generative photo-editing experience on any smartphone, debuting on the Galaxy Z Flip6 and Fold7.

Employee of the Year — awarded Samsung SRIB's highest individual recognition for the impact and quality of the work.

Established the playbook — the interaction patterns and design principles from Photo Assist became the foundation for subsequent AI editing features, including Instruction-Based Editing on the Galaxy S26.

Photo Assist proved users wanted AI editing. The next question: what if users could just describe what they want? Designing a natural-language interface for photo editing that felt intuitive to casual users while being powerful enough for those who wanted precision.

Co-lead Product Designer — shaped the interaction model for instruction-based editing and AI-suggested edits. Extended the design foundation laid by Photo Assist into a more expressive, language-driven paradigm. Also contributed to the broader "Capture → View → Edit → Manage → Share" experience strategy.

Shipped on Galaxy S26. Remaining features slated for Fold8 — a multi-cycle product shaped from the ground up.

Two distinct challenges, one shared principle: make AI generation feel personal, not generic. For Portrait AI, the core problem was style transfer that doesn't destroy what makes a face recognisable. For Sketch-to-Image, turning rough drawings into polished visuals without making the user feel like the AI did all the work.

Lead Product Designer — designed the prompt systems for Portrait AI, scaling across 12+ visual styles while preserving facial identity. For Sketch-to-Image, designed the drawing-to-generation flow.

Selected as one of a small number of Bengaluru designers for an on-site residency at Samsung UX HQ in Seoul to co-develop both features directly with the research team.

Also developed a cross-cultural design framework for AI face features — addressing global facial markers and cultural sensitivity — adopted by the Seoul research team as standard protocol.

Embedding a third-party experience (Snapchat Lenses) inside Samsung's native camera without it feeling foreign. The A-series audience skews younger — the experience had to feel native to Samsung while leveraging Snapchat's creative engine.

Designed the integration UX and established safe-zone guidelines for lens qualification. Proposed two features — Favoriting Lenses and Personalised Lenses — both adopted into the next release cycle.

Surpassed 2 billion cumulative interactions. The two proposed features shipped in the subsequent release, proving the design thinking extended beyond the initial launch.

Samsung Camera ships across S-series, Fold, Flip, Triple Fold, and Tablets — each with different screens, aspect ratios, and hardware. Maintaining a coherent, high-quality camera UI across all of them, over five OS generations, is an exercise in systems thinking.

Owned the full production pipeline: visual specs, GUI/UI documentation, design system guides for A & M-series, testing, and offshore dev handoffs with China. In parallel, redesigned the Magnifier accessibility app across all six form factors and defined the Expert Raw Sky Guide design system — a visual language for Samsung's pro photography app covering planets, stars and constellations.

Three separate innovation explorations: an AI video editing suite aimed at making pro-level video edits accessible, an AI photo editing pipeline that predated Photo Assist, and a Creator's Camera mode designed for content creators who needed quick, polished output.

None shipped as originally conceived — but they weren't failures. The AI photo editing pipeline directly informed the design principles behind Photo Assist. The Creator's Camera exploration shaped Samsung's understanding of their pro-sumer audience. And the video editing work built a foundation of AI interaction patterns the team continues to draw on.