AI Product Engineer,Agent Systems Architect.
I design and build AI agent systems, real-time voice applications, and full-stack products with Next.js and Python, turning product ideas into production-ready software.
Crossing technical boundaries,
powered by systemic architectural thinking.
🛠️ Product Engineering Mindset
Going beyond code to design scalable, commercially viable architectures rooted in product definition, UX, and business logic.
🧠 AI Agent Systems
Building sophisticated agents with deep reasoning and autonomous orchestration using Python, translating cutting-edge AI into practical business tools.
🚀 Full-Stack Delivery
Expertise in React/Vue, cross-platform mobile frameworks, and high-performance Python backends, delivering excellence from blueprint to launch.
AI Workplace Coach, the start of evolution.
Workplace training system based on autonomous agent logic. More than just dialogue, it's a deep reshaping of professional skills.
"Traditional corporate training is costly, slow to feedback, and lacks realistic simulation environments."
How to achieve millisecond-level real-time voice interaction and emotional feedback on high-latency LLMs?
Through a self-developed Agent orchestration layer, response speed increased by 40%, helping companies reduce training costs by 30%.
Reshaping the architecture of AI applications.
Closed-loop framework built for AI SaaS. From model routing to real-time voice orchestration, making intelligence accessible.
"AI developers often struggle with multi-model switching, unstable streaming, and uncontrollable costs."
Design a multi-model dynamic routing gateway that supports millions of concurrencies and has automatic fallback strategies.
Achieved extremely stable streaming delivery, reduced API costs by 25%, and became the foundation for multiple AI projects.
ScrapBear
Full-link Smart Logistics Recycling.
Cross-platform enterprise mobile solution. Transforming complex recycling scheduling into an extremely smooth user interaction and order-taking experience through a full-link service loop.
"Traditional recycling chains are chaotic, with low order matching efficiency and poor user trust."
How to ensure real-time synchronization of logistics tracks in weak network environments and solve distributed lock issues in high-concurrency order grabbing?
Built a standardized recycling delivery flow, increased order processing efficiency by 200%, and broke single-month GMV records.
Content management, simplified.
Designed for high-performance enterprise applications. Achieving peak content delivery efficiency through dynamic resource management.
"Enterprise global expansion led to a surge in content volume, with traditional CMS facing bottlenecks in multi-language and multi-terminal distribution."
How to design a highly available architecture supporting tens of millions of data points and achieve millisecond-level cross-regional content rendering?
Introduced edge computing and Incremental Static Regeneration (ISR), reducing global content first-paint time by 60% and supporting tens of millions of daily PVs.
Your desktop AI companion, where company meets productivity.
A nurturing desktop AI companion. On the surface, a character-driven interactive pet; at its core, an intelligent assistant that understands natural language and safely executes desktop tasks.
"Existing AI assistants are either too tool-oriented and lack emotional connection, or purely entertainment-focused without real utility. Users struggle to find a unified experience of companionship and productivity."
How to enable AI agents to understand user intent and autonomously execute cross-app desktop operations while ensuring system security?
Project in development, expected launch in June 2026. The goal is to create a desktop companion with both warmth and capability.
SKILLS
AI CORE
Core for Intelligent Reasoning
UI ENGINE
React · Vue · Next.js · TS
MOBILE
RN
Uni-app
Mini-Pro
INFRA
FastAPI / Postgres / Docker
FAQ
I am a strong fit for teams building AI products, agent workflows, real-time voice features, internal tools, or full-stack platforms that need both product thinking and hands-on engineering. If you need someone who can move from idea to working system quickly, I can help.
Yes. I can help turn an early concept into a clear product scope, technical plan, and working MVP. This is usually the fastest path for founders or teams that need to validate an AI product before scaling it.
Yes. I work across the full stack, including product architecture, AI workflows, backend systems, and modern frontend delivery. That makes it easier to ship faster without losing consistency between product goals and implementation.
Yes. I can audit an existing codebase, improve architecture, fix delivery bottlenecks, and rebuild weak parts when necessary. This works well for teams that already launched something but need better performance, scalability, or AI capability.
Most projects start with a short discovery conversation around goals, product scope, timeline, and technical constraints. From there, I typically define the right delivery path, whether that means a rapid MVP, a focused system upgrade, or a full product build.
From Definition to Implementation,
Architecting Products with a Digital Soul.
Serving high-velocity startups and SMEs seeking AI transformation. Specializing in rapid 0-to-1 MVP construction, AI agent architectures, and high-performance full-stack delivery. Currently available to start as early as next week.
"