AI in.
Advantage out.

Software and fintech are being redrawn by AI. I'm Xianguan Li, founder of Xianguan-AI — I help enterprises turn AI into an edge, building the agentic workflows that make platforms stand out. Everything below is live, built and deployed end to end.

See the prototypes → Work with me →

The prototypes

● Testing

Talent Bridge

AI enablement, applied to recruiting. Multiple AI agents work together at every step — profile building, intelligent job search, customized resumes, AI-assisted applications, and a recruiter chat with your digital twin.

● In development
Money Tree logo

Money Tree

Wealth management for people without a finance background. Ask about rebalance impact, risk and performance analysis, tax-optimization opportunities, and more.

○ In design

Quantability

A quantitative lens on long-term wealth strategy — not day-trading. Personalized guidance and hints when opportunities present themselves.

Under the hood

Agentic AI, end to end.

Talent Bridge is the proof — a multi-agent system that runs the whole loop (search, judge, write, apply) and represents you to recruiters, with you in control. Not a chatbot with a job board bolted on.

Search

Autonomous multi-agent search

A crew of agents searches boards and company career sites, reads each posting with an LLM, and scores the fit against your profile — not keyword matching.

Generate

A resume written per posting

A writer agent drafts a fresh, tailored resume from your full profile and the specific job — then renders the PDF.

Apply

Human-in-the-loop, via MCP

An MCP server runs on your own machine, so Claude fills the real application form in a real browser — you watch and approve every step.

Represent

A digital twin for recruiters

Recruiters chat with an AI twin of you that answers on your professional background. When a question isn't in your profile, it pushes you a notification to answer in real time — no dead ends.

Built solo · Claude Code

Cloud-hosted, provider-flexible

Six services and a React front end, multi-agent orchestration on hand-built k3s + AWS — designed and shipped solo, with Claude Code as the pair. Swappable model providers — OpenAI, Anthropic, and more. No vendor platform.

Who’s building this — and why

Xianguan Li

Xianguan Li

Enthusiastic Technologist · Experienced Leader · AI / agentic engineer · builds and ships end to end

AI is redrawing the map. Software and SaaS firms are back near the same starting line — and what separates the winners next is how deeply AI runs through their day-to-day operations. AI enablement and AI transformation are now front and center. I’m founding Xianguan-AI to help enterprises seize what AI makes possible — building the agentic workflows that let their platforms stand out.

What Xianguan-AI does

  • Turn entrepreneurial ideas into working prototypes
  • AI enablement — strategy and hands-on practice
  • AI transformation consulting for fintech and wealth-management firms

A technologist and leader with 25+ years across wealth management, securities trading, banking, medical, media, and real estate — leading cloud modernizations that make classic systems more reliable, scalable, and observable, and building the Agile, distributed teams that ship them. I’m an AWS solution architect with deep wealth-management expertise (portfolio accounting, investment and model management, and trading), and an enthusiastic technologist set on using AI to make teams faster and products smarter — hands-on with Claude Code, agentic workflows, AWS, Python, Terraform, Kubernetes, and Snowflake. The prototypes above are that conviction in code, built and deployed solo. I’m open to permanent roles, contract work, and partnerships where AI can make a real difference.

Got an idea?
Pitch it.

If it's picked, we build it together and it joins this page.

Connect

Open to permanent roles and contract work — or a conversation about any prototype. Drop me a line below.

Goes straight to my inbox. No mailing lists, no spam back. Spam-protected by Cloudflare Turnstile — no cookies, no tracking.

Got it — thanks! I read every message and reply when there's something worth saying.

github.com/xli-26