Winkie Tian is a product designer At the intersection of Landscape Architecture and technology.

Cambridge, MA, United State
Harvard Graduate School of Design MLA '26
scroll

Draw With Music

A drawing app practice that explores drawing with music through colorful brushes controlled by human-screen interactions. There are two modes: you can draw using your mouse or gestures with or without guidelines. You can also track and play back your creations.

draw to Play

A "Demo Scene" exploration of Boid algorithm visualization - a drawing canvas, a music game, and an autoplay demo inspired by Revision 2024. It explores music analysis through tone.js combined with particle motion control through p5.js.

Image To Music

An AI music generator works by simply uploading a picture and moving the sliders to choose sound vibrance, mood, and style. Through double API calls, the atmosphere and vibe of the picture are understood and processed by the large language model to create the music you want.

Field-based fabrication

This project explores the use of field-based geometry generation and scalar data mapping as a framework. The result is a voxel-based field of cubes - a minimalist representation of a continuous scalar field (x-coordinates) visualized discretely through form and color. The output is intended for multi-color 3D printing that retains both geometric and chromatic information.

The Lost Rhythm of Life

PRODUCED BY Winkie Tian | xuanxuan liu | beining wu

2025 @MIT AI FILMHACK

Inspired by the theme DANCE IN LIFE, we explored the boundaries of generative AI in animation video production by integrating dance movement with music. Our experiment involved various AI tools, models, and technologies, focusing on testing their capability in motion tracking, camera movement analysis, ambient emotion expression, and understanding music videos as a narrative medium for content creators.
 
The first clip narrates a protagonist who loses her ability to dance after an accident and withdraws from life itself, while her personal AI bot brought her to the virtual world. We also integrate AI-composed music and voice into this piece to construct dialogues and interactions.

The second clip shows the transition between the two worlds. In the virtual world, dance becomes limitless—she floats through water, leaps across lava, and bends reality with movement. Yet, this freedom is fleeting; she realizes that true dance, like life, must exist in imperfection and presence. Her journey mirrors personal transformation: falling, rebuilding, and embracing life anew. Life, like dance, is not about perfection but about movement, resilience, and expression.

Reimaging Sustainable Energy Use with AI Emergence:
Green Decentralized Data Centers for the Publics

The conventional GPU clusters for training AI usually cost billion for constructing a single campus and requires Enormous Power Supplies. These enclosed big boxes are often invested by big tech giants, occupying large urban fabrics with impermeable surfaces. However, technological development should be considered alongside environmental sustainability.

This project proposes green, decentralized data centers for public environmental monitoring and restoration. By transforming gray infrastructure into phyto-remediation landscapes supported by IoT technology, residents actively participate in planting, water sampling, and data tracking through the My Green Eyes app. Five species of plants are selected to remove pollutants, and solar-powered, portable data units are installed to automate data entry and analysis. Recycled materials from dismantled parking lots and warehouse buildings are repurposed to form a sustainable, circular economic system of regional material flows.

If we understand data centers as middleware transforming energy into readable data, we can assume that with temporary, portable, and transformative energy units, data centers can be programmed differently to permanent structures, with public participation and cyclic planting schemes incorporated into its construction and maintenance. This approach integrates public participation and education, enabling communities to take ownership of environmental data and drive collective decisions for groundwater restoration and ecological resilience.

See full Project Description

Generative aI in Art, design, and urban landscapes

TUTORIAL SERIES 2025 Winter j-term  | @Harvard gsd

I initiated and led teaching a Stable Diffusion curriculum in bridging generative AI into spatial design as a medium to explore form generation and design representation. Through four-day lecture series and hands-on workshops, students were introduced to LLM, GAN and computer vision theories, LoRA training, ControlNET, AnimateDiff, and ComfyUI. I also invited guest speakers to share practical insight, such as the cutting-edge application of generative workflows in design industries.

Students were encouraged to explore the creative role of AI, engaging and augmenting their knowledge and expertise in architecture and urbanism. This experience enabled all of us to expand the creative spatial possibilities of AI tools with a critical reflection on their limitations and challenges.
(01)

HarvardXR Conference Lead

Sep 2024 - Now

As the Design & Marketing Team lead, I coordinated the 2025 'The Augmented Self' XR Conference. I led the design of the branding visual system across digital and physical assets and managed team resources for social media & press marketing, tabling, outreaching, logistics, and onsite operations to ensure smooth conference experiences for all speakers, sponsors, volunteers, and audience.

(02)

Product Design Internship

06/2024 - 08/2024 - Gustouch Co. Ltd

I designed prototypes, use cases, and user experiences for AI experiments at a B2B SaaS company. I mocked up new text-to-image features for the clients to customize branding IP visuals, providing dataset and directional support for embedding generative AI functions into the existing web application.

(03)

Product Design Internship

05/2024 - 06/2024 - Metrodata Co.Ltd

I prototyped and mocked up an AI data visualization dashboard for a property services company with the technologies of RESTful APIs, RAG, embeddings, and prompt engineering to ensure the accurate delivery of data with clear visual identity systems.

See project