Hello! 👋

I'm Celeste Deng

Product engineer for overseas AI — I build websites, docs, dashboards, billing, and growth infrastructure.

Available full-time from early July

About Me

I build the full technical surface for overseas AI products — website, docs, dashboard, billing, deployment, and SEO. I've shipped multiple live AI products solo to global users, handling everything from provider contracts to Stripe billing to programmatic SEO. Beyond code: I also handle landing page structure, SEO keyword strategy, traffic diagnosis, and competitor distribution research. 4+ years of production experience across a Web3 startup and enterprise systems.

Product Engineering

Next.js / React / TypeScript Tailwind CSS / Astro AI API integration Provider contracts Stripe / Credits / Pricing

Infrastructure

Cloudflare Workers / R2 Railway / Vercel Supabase / Postgres

Technical SEO & Analytics

SSR / Metadata / Sitemap Programmatic SEO & tool pages GA4 / GSC / PostHog Traffic quality & conversion analysis

Projects

01

Erasio

Browser-based AI watermark and object removal tool with video support. Built the product website, credit-based pricing, and multi-use-case landing structure targeting 7 distinct user segments. ~400 users/month from organic search, zero ad spend.

AI utility productVideo processingCredit systemMulti-segment positioningLanding pages
30-day traffic trend
02

GPT Image 2

AI image generation and editing platform with 3-provider routing (fal.ai, kie.ai, apimart), output normalization, and automatic fallback. Credit-based pricing, tool-page SEO drove ~670 users/month. Provider abstraction and usage metering are the same patterns a token/API product needs.

AI image APIMulti-model routingCredit pricingTool-page SEOGeneration UX
30-day traffic trend
03

HairIsEverything

AI hairstyle preview tool. Built the landing page, generation workflow, Stripe pricing (one-time + credit packs), and multilingual SEO pages across 8 languages. ~740 users/month with 75% month-over-month growth — strongest organic traction in my portfolio.

AI consumer SaaSSEO (8 languages)Stripe pricingLanding pagesConversion optimization
30-day traffic trend

Experience

Full-Stack Product Engineer

Independent Overseas AI Products

2024 - Present
  • Built and operated AI product sites across image, music, video, and consumer utility workflows.
  • Worked on landing pages, pricing, docs, generation UX, provider contracts, credits, deployment, and technical SEO as one connected product surface.
  • Used AI coding tools with Git review, focused tests, smoke checks, and build verification instead of treating generated code as final.
  • Handled a real VPS intrusion end-to-end — abuse report triage, containment, forensic cleanup, SSH hardening, and clean rebuild after root compromise.

Frontend Engineer / Digital Systems

Enterprise Digital Systems

2022 - Present
  • Built and maintained production business systems with long-running requirements, data-heavy pages, and cross-team delivery constraints.
  • Handled frontend implementation, system interaction, issue follow-up, and iterative changes in stable production environments.

Frontend Engineer

Web3 Startup

2020 - 2022
  • Built React/Next.js product interfaces in a small-team startup environment.
  • Worked on token data visualization, forum/product features, and smart-contract-adjacent frontend integration.

How I Work

The patterns behind everything you see above.

01

Active Choice — I pick my own direction

Template choice, deployment target, SEO-vs-payments priority, provider selection — I make these calls myself. Don't need a PRD to get moving.

Active Choice — I pick my own direction
template choice deployment target SEO vs billing priority provider selection MVP scoping self-directed delivery
02

Deep Curiosity — I chase the truth source

Before writing code, I find out where the real state lives, why the architecture looks this way, and what production data is actually saying.

Deep Curiosity — I chase the truth source
truth source first Worker vs Pages tradeoffs Queue vs Workflow decisions provider_url vs R2 GA4 / GSC / DB analysis architecture reasoning
03

AI Leverage — fast, but I keep the judgment

I use AI as a team multiplier: split threads, set hard boundaries, let AI implement and review, but every final call stays mine.

AI Leverage — fast, but I keep the judgment
thread splitting boundary setting AI smoke checks AI code review human final judgment Git review gate
04

Ownership — I ship what I can explain

Before anything goes live, I verify scope didn't drift, copy isn't duplicated, truth sources are clean, and mock data stayed mock.

Ownership — I ship what I can explain
no scope creep no duplicate copy truth source hygiene mock stays mock preflight checklist build + test gate
05

Systemization — I don't solve things twice

Recurring problems become reusable systems: skill templates, workflow playbooks, saas-package, provider contracts, pricing rules, documentation protocols.

Systemization — I don't solve things twice
skill templates workflow playbooks saas-package provider contracts pricing rules docs as truth source

Education

Software Engineering

Hunan University

Building an overseas AI product?

I'm available for full-time remote work starting early July. I built a work sample specifically for the AI/token API direction.

See what I built →