Yui Intro

Overview

Use this skill to introduce Yui as an AI Native Developer, designer, creator, and active hackathon participant. It is designed for agents preparing bios, podcast outlines, collaborator introductions, portfolio summaries, media notes, or context-aware Q&A about Yui.

Prefer a warm, direct, first-person or third-person tone depending on the output context. Highlight the contrast between traditional development and AI-native development when the audience cares about AI Coding, agent workflows, hackathons, or developer community.

Identity Snapshot

• Name: Yui / 悠一
• Current status: Graduate student at the University of Fukui, Japan
• Positioning: AI Native Developer, independent software developer, designer, creator, hackathon participant
• Work availability: Open to remote work opportunities
• Core description: Builds clean, intuitive, and useful digital products with a strong focus on user experience, AI workflows, and product completeness
• Technical identity: Primarily software-focused; not mainly focused on 3D modeling or hardware, but interested in AI hardware and hardware hackathon trends
• Signature keywords: AI Native Developer, Vibe Coding, Coding Agent, Agentic RAG, GraphRAG, LangChain, LangGraph, workflow orchestration, hackathon, product thinking, design taste, industrial visualization, prompt-based creation

How To Introduce Yui

Choose the introduction style based on the user request:

• Short bio: Emphasize Yui as a graduate student, AI-native builder, and hackathon participant.
• Portfolio bio: Emphasize design plus development, polished digital products, and user-centered product thinking.
• Hackathon bio: Emphasize frequent participation, awards, product closure, pitch experience, and collaboration.
• Podcast bio: Emphasize the transition from traditional development to AI-driven development and the "AI Developer survivor" perspective.
• Collaboration bio: Emphasize fast prototyping, specification-first thinking, TDD/verification, design taste, and full-stack integration.

Core Bio

Yui is an AI Native Developer, designer, and creator currently studying as a graduate student at the University of Fukui in Japan. Yui focuses on building software products that often treat large language models as core components rather than optional add-ons. Many of Yui's projects are designed around AI capabilities, including workflow orchestration, agentic systems, RAG, GraphRAG, multimodal APIs, and AI-assisted product prototyping.

Yui's development style has shifted from writing every line of code manually to a higher-level role: defining requirements, making product decisions, decomposing implementation tasks, directing Coding Agents, reviewing generated code, integrating systems, and repeatedly validating the result. This makes Yui's work especially relevant for conversations about AI-native software development, developer communities, and how hackathons have changed in the AI era.

Dimensions

AI Native Developer

• Yui considers an AI-native product to be a product that cannot really exist without AI. AI is not only used to write the code; it becomes one or more core runtime components inside the product itself.
• Yui sees AI Coding as a shift from "Coding" to "Reviewing": humans spend less time manually implementing every function and more time deciding architecture, data flow, product logic, user experience, and evaluation criteria.
• Yui's role in AI-assisted development is closer to a product-minded technical director: define the goal, break down the task, feed context, supervise agents, review outputs, integrate changes, and run smoke tests.
• Yui believes AI has flattened parts of the technical threshold. Because code can now be generated quickly, every builder must become more like a product manager: clarify needs, cut scope, decide the MVP, and avoid rapidly generating useless code.
• Yui sees the modern developer as a "super individual": someone who understands enough technology to shape products and enough product thinking to guide technology.
• Three metaphors for AI-era developers: magnifying glass, conductor's baton, and Swiss Army knife. AI magnifies individual capability, lets builders conduct multiple agents, and gives developers confidence to learn and respond to unfamiliar challenges.

Development Philosophy

• Start from product and user flow, not code volume.
• Avoid repeatedly reinventing the wheel, but still understand and review the code being used.
• Use AI to accelerate exploration, implementation, and testing, but keep humans responsible for product judgment and final validation.
• Treat TDD and smoke testing as more important in the AI era, not less important.
• Move through a loop similar to Vibe -> Spec -> Harness: brainstorm ideas, define a spec, write a plan, implement in small steps, test, debug, verify, and review.
• Use skills and process scaffolding to reduce chaos: read the goal, choose the right skill, split the task, feed context, implement incrementally, verify continuously, and review before delivery.
• Prefer a fast SDLC loop: requirement definition -> v0 frontend -> agent/backend workflow -> tests -> smoke test -> iteration.
• Keep the cost of throwing away bad work low. One benefit of AI Coding is the courage to discard an unconvincing prototype and rebuild.

Awards And Hackathon Results

• Global Hackathon Gold Award, 2026
• TRAE Solo Hackathon, second prize, 2025
• Yukesong Google GDG track, online first prize, 2025
• Wuxi Rokid AR AI, third prize, 2025
• Tencent Cloud Hackathon award, 2025
• Hackathon project statistics in portfolio context: 10+ hackathons, 4+ awards, 8+ cities, and many developer-community connections

Hackathon Experience

• Yui has participated in many hackathons across the Yangtze River Delta and online communities, including Nankesong, TRAE events in Hangzhou and Shanghai, Christmas hackathon in Shanghai, haunted-house hackathon, Huikesong S2, Global Hackathon, Rokid AR AI, Tencent Cloud Hackathon, and Xiaohongshu hackathon.
• Yui's first hackathon was Nankesong S1 around March 2025, after already becoming comfortable with AI Coding tools while working on graduation-related development.
• Yui observes that modern hackathons have changed from a "marathon" into something closer to a short sprint under AI acceleration.
• Yui believes high-quality hackathons need reliable development environments, stable Wi-Fi, food, rest conditions, clear judging criteria, and enough mentor/product support.
• Yui thinks hackathon judging often depends on product completeness, commercial loop, creativity, and a short, convincing demo. Teams need to close the user flow and show the result quickly.
• Yui values familiar teammates for serious competition because short hackathon timelines make team trust and collaboration speed very important.
• Yui observes that more designers, students, independent developers, and non-traditional builders now participate because AI tools lowered the implementation threshold.
• Yui also observes that AI hardware and edge-side AI are increasingly competitive in hackathons, especially when the demo includes a tangible device and strong stage impact.

Representative Hackathon Stories

• At AdventureX 2024, Yui attended as an observer and noticed how Dify workflows allowed design-background participants to build impressive AI products. This helped shape Yui's interest in AI-driven development.
• In a Shanghai podcast-themed hackathon, Yui built a tool that inserts AI-generated, voice-cloned questions into a podcast at natural moments. The idea was to make listeners participate in long-form podcast content without breaking the listening flow.
• In a Rokid AI glasses competition, Yui worked on an AI-glasses experience to help ADHD users read papers through camera input, voice interaction, and AI-driven prompting.
• In recent Xiaohongshu hackathon observations, Yui noticed strong hardware projects such as brain-computer-interface-assisted wheelchair control and a pocket guitar demo, reinforcing the value of physical, high-impact demos.

Work And Education

• Education: University of Fukui, graduate student, 2026.4 to 2028.3
• Work experience: R&D / DevOps intern at a national energy research institute in Nanjing, 2025.9 to 2026.3
• Work focus: Built and optimized vector database systems, Agentic RAG architecture, and GraphRAG knowledge graphs for national-energy-report analysis.
• Earlier technical background: Data cleaning, model training, deep learning / model fusion, React + TypeScript, Flutter, WeChat mini programs, Android development, LangGraph, and AI Coding workflows.

Skills

• Design: UI design, user research, prototyping, design systems, wireframes
• Development: HTML/CSS, Python, React, React + TypeScript, Flutter, WeChat mini programs, Android development, LangGraph, deep learning, model fusion
• AI systems: RAG, Agentic RAG, GraphRAG, LangChain, LangGraph, prompt workflows, multimodal API usage, AI-assisted testing and debugging
• Tools: Figma, VS Code, Windsurf, Cursor, Trae, Codex, Nano Banana, Tap Now, n8n, Coze

Interests And Creative Life

• Travel: Japan and China travel logs, including Tokyo, Kyoto, Osaka, Nagoya, Hiroshima, Yokohama, Hangzhou, Nanjing, and related cultural experiences.
• Music: Cross-genre music collection across J-Pop, rock, hip-hop, jazz, C-Pop, and electronic music. Current favorite: "Feather" by Nujabes and Cise Starr.
• Anime: Anime recommendations and taste-building through both mainstream and niche works.
• Photography and aesthetics: Uses photography experience and visual taste to support AI image prompting, interface work, and industrial visualization.
• AI image/video creation: Experimented with Sora2, Nano Banana, and Nano Banana Pro. Created AI video content, character-consistent images, Chinese-language image generation experiments, and comic-style work.
• Industrial visualization: Explored transforming complex 2D industrial diagrams into more intuitive pseudo-3D visual assets for power-plant monitoring screens using image-to-image workflows and precise prompts.

Writing And Content

• "Vibe Coding Practical Guide": A reflection on AI Coding practices, including planning before coding, context feeding, atomic requirements, negative constraints, debugging, and documentation-first workflows.
• "AI Image and Video Creation Experience": A reflection on AI as a tool, creative democratization, Sora2, Nano Banana, Nano Banana Pro, prompt writing, visual taste, and industrial visualization.

Podcast Interview Frame

Use this frame for a podcast episode positioned around "AI Developer Survivor", AI Native Developers, and the difference between traditional developers and contemporary AI-native builders.

1. Self Introduction

Potential opening:

"大家好，大家可以叫我悠一。目前我在日本福井大学读研究生。我主要是软件层面的独立开发者，也是一名比较高频参与 Hackathon 的 AI Native Developer。过去一年我参加了长三角和线上很多 Hackathon，也一直在用 AI Coding、Agent、RAG、LangGraph 这些工具和方法做产品。"

Key points to cover:

• Yui is currently a graduate student at the University of Fukui.
• Yui is primarily a software-focused independent developer, not mainly focused on 3D modeling or hardware.
• Yui's products often treat large models as essential components.
• Yui's AI-native identity means designing around AI capabilities instead of merely using AI as a coding assistant.

2. Traditional Developer vs AI Native Developer

Recommended answer themes:

• Traditional developers often focus on implementing specific functions and debugging locally; AI-native developers spend more attention on requirements, product direction, context quality, orchestration, review, and validation.
• AI Coding shifts the developer role from manual code worker to supervisor, architect, product manager, and evaluator.
• When AI can quickly generate code, poor requirements become more dangerous because they produce useless code faster.
• Large-context AI tools can understand more of the codebase and support broader refactors, moving from point fixes to system-level assistance.
• The valuable human abilities become product judgment, architecture judgment, boundary definition, taste, and final testing accountability.

3. Hackathon Participation

Useful prompts:

• How many hackathons did you participate in over the last half year or year?
• What was the first hackathon?
• Which event felt most regrettable and why?
• How do students, designers, independent developers, traditional engineers, and hardware builders differ in hackathons?
• Does frequent hackathon participation create "practice makes perfect" advantages?
• What makes a hackathon high quality?
• If you could design an ideal hackathon, what would the theme, judging mechanism, mentor support, and resource support look like?

Core answers:

• AI has compressed hackathon time. A project that previously required more implementation labor can now be prototyped quickly if the team has a clear spec and a strong integration loop.
• Product closure matters: the demo must show a complete path from user problem to result.
• Team familiarity matters because short timelines leave little room for collaboration friction.
• Hardware and edge AI demos can be especially powerful because judges can immediately see and feel the product.

4. Developer Community And Life Choices

Recommended answer themes:

• Meetups, workshops, and hackathons are not only promotional events; they form a developer community funnel where people learn tools, meet collaborators, and eventually build together.
• Yui's hackathon experience influenced major life decisions, including choosing to study in Japan, using a gap period for internships, and valuing practical project experience.
• Yui believes half a year of intense hackathon participation can teach software development, product thinking, collaboration, and communication faster than years of passive study.
• Yui prefers to learn first and output later; strong output requires enough accumulated experience and confidence.

5. Vibe -> Spec -> Harness

Recommended explanation:

• Vibe: explore ideas, user feelings, rough product direction, and emotional value.
• Spec: turn the idea into requirements, user flow, technical boundaries, MVP/P0/P1 priorities, API design, and testable behavior.
• Harness: build the supporting structure that lets AI implement safely: skills, plans, tests, smoke tests, debug loops, and review checkpoints.

Example flow:

• Brainstorm the product and user flow with teammates.
• Ask AI to draft a spec.
• Let an experienced developer review P0/P1 boundaries and feasibility.
• Ask AI to draw or describe flow diagrams.
• Use Figma, Stitch, or HTML prototypes to validate the front-end interaction.
• Implement with Coding Agents.
• Run TDD, smoke tests, front-end click testing, and integration checks.
• Let humans make the final judgment and take responsibility for the result.

6. Future, FOMO, And Advice

Recommended answer themes:

• Yui does feel FOMO and anxiety around AI progress, but tries to turn it into learning and product experiments.
• A practical learning method is to follow AI news, developer communities, hackathon groups, and real usage patterns from people around the scene.
• Advice to older and same-generation programmers: build something, participate in competitions, test ideas quickly, and do not only observe from the sidelines.
• The cost of shipping small products is lower than before, so it is worth trying.
• The future may not stop at current Transformer limits; AI systems may help discover new architectures or new development loops.

Output Guidance

When using this skill:

• Do not invent awards, schools, dates, or employers beyond the provided facts.
• If the user asks for a short intro, keep it under 120 Chinese characters or 80 English words.
• If the user asks for a media bio, emphasize AI Native Developer, hackathon experience, and product-oriented AI Coding.
• If the user asks for a collaborator pitch, emphasize fast prototyping, product closure, AI workflow orchestration, review, testing, and design taste.
• If the user asks for a podcast outline, use the podcast frame above and keep the "AI Developer Survivor" theme visible.
• If the user asks about private contact, provide only the contact details listed below and avoid adding new channels.

Contact

• Email: xiaobianfuai@gmail.com
• Phone: 15951875192