Full-Stack B2B Marketer.
I Also Run Marketing on AI.
Paid and inbound, lifecycle and CRM, and the operations layer under them, run with AI on the team: agents carry the repeatable work, people keep the judgment. At PingPong I owned acquisition and lifecycle on a payments line as its annual revenue grew from $2M to $250M+. The tools I could not buy I built, and every one is on a live URL.
- Full funnel, paid to retention
- GTM systems, CRM, attribution
- Lifecycle and email systems
- SEO, GEO and content engines
- AI agents that run the above
U.S. Bank
Gusto
TikTok
Clover
Jeeves
PingPong
BlackwallU.S. BankGustoTikTokCloverJeevesPingPongBlackwall
Owned the demand side of one cross-border payments line at PingPong.
Paid acquisition and lifecycle, end to end, from 2021 to 2025: first as Marketing Manager, Ads & Lifecycle, then as Senior Digital Marketing Manager, Growth & Ops. A regulated cross-border product, so speed could never cost us the audit trail.
- $2M → $76M in year one, past $250M+ by year two. Annual revenue through the line: inbound in year one, then an account-based motion built with sales in year two.
- LinkedIn cost per lead $300+ → ~$160. Google registrations $90+ → ~$35, with volume held through both.
Which role are you hiring for?
Enterprise ABM, international launches, lifecycle growth, and the sales partnership around them. I ran all four as one funnel, with numbers that reconcile to finance. Often posted as Senior Growth Marketing Manager, B2B or Demand Gen Manager.
Scope: the demand side of the B2B cross-border payments line at PingPong (2021–2025), paid and lifecycle end to end, as annual revenue grew from $2M to $250M+, with AI now writing the reports. Earlier, VP of Marketing at ToLocal (2016–2021), a team of 12 across the US and China on a $10M annual media budget at 80%+ ROI.
Agent-native go-to-market: multi-model agent pipelines research, enrich, build the campaign and report, while a person sets direction. Often posted as GTM Engineer, AI GTM Lead or MarTech Lead.
Scope: nine AI GTM tools designed and shipped solo, every one on a live public URL. Currently Senior Marketing Manager at ThinkingAI, leading GTM systems and marketing ops.
I architect workflows, write routing logic, keep attribution clean and govern the site. I build the systems every other team reads its numbers from, so sales, marketing and finance open the same figure on the same morning. Often posted as Senior or Lead Marketing Operations Manager, or Marketing Technology Manager.
Scope: Senior Digital Marketing Manager, Growth & Ops at PingPong (2023–2025), sole owner of a governed 8-tool GTM stack and all of Marketing for 6+ months.
I run the programs and
build the AI that runs them
I'm Daniel Liu, a marketer who wandered too close to operations and campaigns, discovered AI automation along the way, and decided to stay.
A decade later that detour is the whole job: I sit where GTM judgment meets hands-on technical execution, connecting campaigns, CRM, tracking, lifecycle, reporting and AI automation into one revenue engine a CFO can audit.
- ToLocal, 2015–2021. Founding marketer, then VP of Marketing from 2016, leading a team of 12 across the US and China on a $10M annual media budget at 80%+ ROI.
- PingPong, 2021–2025. Marketing Manager, Ads & Lifecycle, then Senior Digital Marketing Manager, Growth & Ops, on the cross-border payments line that went $2M to $250M+.
- ThinkingAI, 2026 to now. Senior Marketing Manager.
Six workstreams,
one AI-native growth system
I owned all six end to end, in the stack and the regions they ran in, and rebuilt each one around the part a model can now carry.
I ran paid acquisition and lifecycle on the cross-border payments line that went from $2M to $250M+ over two years, across NA, EU and SEA.
Learn more → 02 User AcquisitionPaid across Google, Meta, LinkedIn and programmatic. LinkedIn CPL $300+ → ~$160, Google registrations $90+ → ~$35. Earlier, a $10M annual media buy at ToLocal.
Learn more → 03 Email & LifecycleI build behavior-triggered journeys from signup to retention, keyed to what the account has done. Identity verification (KYC) submissions climbed from roughly 10% to 35–40%.
Learn more → 04 SEO / GEO & ContentAt PingPong, 2023–2025, I built the editorial workflow that lifts a draft from a blank NeuronWriter score, a 0–100 measure of how well a page covers the terms it targets, into the 70s–80s. I run technical SEO, and those pages have to surface when someone asks an AI.
Learn more → 05 Marketing Ops & AnalyticsI cut lead handoff from ~12 hours to 1–2 minutes across 200+ governed workflows, and got the monthly report under five minutes from one to two hours, with a model flagging what needs a human. The stack is Salesforce, HubSpot, 6sense and GA4.
Learn more → 06 Partner & AffiliateI carried six years of B2C affiliate into a B2B fintech program: I set the commission tiers, vetted the partners, wired per-partner attribution and screened for fraud before payout.
Learn more →Working prototypes and one production tool
Every one of these started as a problem slowing down work I was running. I built the fix myself, in code, with AI carrying the part that used to eat the week. All of them are live on a public URL.
Ad copy that clears policy before you export it
React
TypeScript
Vite
Tailwind
Gemini 2.5 Flash
DeepSeek
VercelReads your site for the voice, writes for Google, Meta, LinkedIn and Bing, and checks every line against policy while it writes.
Catch the non-compliant phrase before legal does
ReactTypeScriptViteTailwindGemini FlashVercelPaste a fintech draft, pick the jurisdiction, get every risky phrase quoted against the clause it trips.
Find the policy violation before the platform does
ReactTypeScriptViteTailwindVercel FunctionsIt checks headline, body, URL and creative against ~84 platform clauses and your own rules. You get a safety score with every risk explained, before upload.
Pull a creator's whole catalog, get the transcripts out
ReactTypeScriptVite
Express
Neon PostgresGemini 2.5 Flash
FFmpeg
Playwright
Resend
StripeVercelPoint it at a Douyin or TikTok profile; it comes back with the whole catalog transcribed, translated and in a spreadsheet. It runs on your own session, which is what keeps the account safe.
Applications, resumes and interviews in one workspace
React 19TypeScriptViteExpress 5
PrismaGemini 2.5DeepSeek
GPT-4o · WhisperStripeVercelOne pasted job description becomes a tracked application, a resume run through a four-pass tailoring engine with its ATS read-out, and interview prep built from the same context.
Run client Meta accounts through the API, with a person on every change
React 19TypeScript
Next.js 16
Meta Graph APIOpenAIGeminiVercelAn agent drafts the campaign, the ad sets, the creative and the copy, then reads results back. Anything that touches a live account passes a guardrail, a named approver and a typed confirmation.
See the real tech stack behind any competitor's URL
Next.js 16React 19TypeScriptDeepSeekPlaywright
BrowserlessStripeVercelFingerprints the bundles, reads DNS and email posture, maps subdomains, then opens the site in a real browser.
Read the strategy your competitors put into market
PythonDeepSeekPlaywrightMeta Graph API
Ads TransparencyStripeVercelPulls a rival's whole ad library and returns the campaign anatomy: market, message, funnel posture. Spend stays out, unestimated.
Notes from the engine room
Working numbers and postmortems from the systems on this page.
Playbooks, postmortems and working numbers from ten years of B2B growth, filed under growth, lifecycle, SEO and ops.
Read → 02 An ABM list from technology signalsI built a North America target universe out of the technology a company runs. Six hand-checked filter rounds took precision from 2% to 31%.
Read → 03 Search Ads research as an agent pipelineFour models, each on the part it is good at, cut 3,156 competitor keywords down to a 47-keyword buy list the Google Ads API ships.
Read →In their words
What I'm building now, at ThinkingAI
I run marketing here on my own, and I have been rebuilding the operation on AI: content and social pipelines, the website, the reporting layer, and the ops beneath them. Above are the case studies; here is what is running now. The newsletter ships weekly and opens at about 22%, on a list past a thousand, almost all of it from the signup popup on the site.
One loop, run on
every system I own
Diagnose
Find the real constraint before spending on top of it.
PingPong, 2021: spent nothing new until the tracking could be trustedInstrument
One cleaned dataset everyone reads. Definitions before dashboards.
8 standing reports, one dataset, every team on the same numberAutomate
The judgment stays human. The busywork doesn't.
200+ workflows: lead handoff ~12h → 1–2 minutes ThinkingAI, 2026: the same loop now runs as reusable multi-agent systems →Scale
Add budget only on instrumented ground.
Cross-border payments line: $2M → $76M → $250M+The site itself is a work sample
Google's hardware ran the audit on a dated build, next to the sites this one was modeled on. Each score comes with the engineering decision behind it.
Let's talk about the role
Hiring for AI-native growth, AI GTM, or marketing ops and systems? Tell me about the team, the role, and the operating problems you need solved. I'll tell you honestly whether I'm the right operator for it.
Open to full-time roles in AI-native growth, AI GTM, and marketing ops and systems. Currently Senior Marketing Manager at ThinkingAI, so you get a straight answer either way. Tell me your start window in the form and I'll tell you mine in the reply.