Jude Nwachukwu · application for xyz engineer
vibe-coded, single file
→ a thesis, not a résumé

Everything
I build
is a router.

I'm an AI infrastructure engineer. I build systems that take a signal and route it to the right backend — compute, agents, payments, quantum hardware.

Your XYZ role is the same problem in a different dialect. A prospect signal has to find the right play, the right person, and the right channel, fast, with almost no headcount. You don't need a marketer who learned AI. You need someone who already thinks in pipelines.

12+ tools shipped
24+ open source repos
2,111× IBC speedup
93.6% quantum fidelity
01 — PROOF

Four routers I've already shipped.

Every one of these is the same shape: a signal comes in on one side, gets classified and enriched, and goes out to the cheapest-viable destination. The category changes. The pattern doesn't.

Router · Compute
Edge Router
workload local | cloud-gpu | quantum
Routes AI agent workloads to the cheapest viable backend. Dynamic health checks, automatic failover. Three tiers: local edge (free), cloud GPU, quantum hardware.
in route local gpu qpu github →
Router · Agents
ACCR
agent call sol | avax | stellar | base | ton
The Visa + Cisco layer for the AI agent economy. Multi-chain settlement, oracle service, MCP protocol server, quantum-validated routing using IBC consensus.
in hub sol avax xlm base github →
Router · Payments
SkillPay
usage credit debit builder payout
Stripe for the agent economy. Universal credit-based billing, 3 lines to integrate. Users buy credits, skills charge per call, builders earn.
user meter split pay github →
Router · Hardware
Quantum Bridge
qiskit originir (wukong 72q)
Only Western-built bidirectional transpiler to Chinese quantum hardware. Commercial API with credit billing. 93.6% GHZ-3 fidelity. Nobody else ships this.
qiskit wukong github →
02 — THE FIFTH ROUTER

Here is the one
I'd build for you
on day one.

The JD describes a pipeline: prospect intelligence → enrichment → outreach → CRM. That's a router. Pick a signal below and watch it move through the stack. Every stage is something I can stand up in a week with Claude, a handful of APIs, and a GitHub repo full of markdown.

Stage 01
Signal ingest
awaiting
Stage 02
Enrichment
awaiting
Stage 03
Qualification
awaiting
Stage 04
Outreach + CRM
awaiting
Outputs below are simulated for demo purposes — the real pipeline hits a Claude API, a HubSpot/Salesforce connector, and a tracked-links service. The logic is the same: one router, four stages, green at the end means a booked meeting.
03 — Why me

I spent five years in risk management at State Farm — evaluating complex scenarios, managing claims portfolios, making data-driven calls at scale. Then I left to build AI infrastructure because I could see what was coming and wanted to build the rails, not just use them.

Since then I've shipped twelve tools, opened twenty-four repos, filed a patent on a quantum consensus protocol, and built the only Western bridge to Chinese quantum hardware.

What that profile actually says: I finish things. I am deeply comfortable with ambiguity, terminals, APIs, markdown repos, and systems that didn't exist yesterday. The JD calls for someone who "does two impossible things before noon because nobody told them it was hard." I have been that person, alone, for the last twelve months.


04 — THE JD, MAPPED

What you asked for, already done somewhere on my GitHub.

Every line in your “What You Will Do” is something I've already built an analogue of. Here's the translation table.

JD Requirement
My Translation
Build AI-powered XYZ infrastructure: CRM, automation, analytics
I built ACCR — a full multi-chain settlement stack with oracle, MCP server, and dashboard. Wiring HubSpot is the easy version.
AI workflows for prospect intelligence — profiling, exec moves, job signals, RFPs, news
Same shape as Edge Router. Ingest → classify → enrich → route. Swap compute for prospects and the architecture is identical.
Sales enablement: transcript pipelines, AI briefs, proposal drafts, self-service content tools
Agent Launchpad already does this for developers — natural language in, deployable package out. Retarget it at the XYZ team.
Curate XYZ repository (likely markdown in GitHub)
I run twenty-four of those. My instinct is already “put it in the repo, write a README, make it queryable.”
Vibe-coded interactive experiences for marketing and thought leadership
You're reading one. Single file, no framework, zero dependencies besides two fonts.
Track XYZ metrics: pipeline, ABM scorecard, funnel, campaign performance
Built The Pulse — a real-time dashboard with SQLite backend and React frontend. Same skeleton, new schema.
Mission-aligned — climate, energy, hard problems
Quantum hardware isn't a hobby for me. It's the bet that critical infrastructure gets rebuilt this decade, and I want to work where that's the point.