The growth engineers worth following.

Workflow

Built a Claude Code 'Research Agent' that ingests hundreds of AI-moderated interview transcripts and, in one session, finds patterns, tests significance, layers in competitor context, and auto-generates charts, slides, and reels for messaging strategy.

Workflow

Built Vercel's GTM 'Lead Agent': inbound forms get auto-enriched, deep-researched, and AI-qualified into buckets, then posted to Slack with reasoning and pre-drafted replies, moving inbound SDRs from research-and-write to review-and-send and cutting that headcount from roughly ten to one.

Workflow

Runs almost his entire GTM data-science practice inside Claude Code for high-agency exploratory analysis, weighing its flexibility against Clay’s reliability guardrails.

Workflow

Built ElevenLabs' AI marketing stack, including a Cursor-coded website-localization pipeline (GitHub Actions firing a custom LLM translation server with per-language brand prompts) that replaced a $40K/yr SaaS tool plus six figures of agency spend, a Granola-to-GPT case-study generator, and a WhatsApp-to-Claude MCP.

Workflow

Breaks down how Clay automates its own inbound: a form submission flows Typeform → Clay → HubSpot → Slack to enrich, score, qualify, personalize, and fire a first touch before a rep steps in.

Workflow

Architected Superpower's waitlist engine: a tiered referral mechanic (one referral jumps you to the top 5,000, two to the top 500) plus shareable 'Bio Age' health cards and creative drops, driving 150K+ signups in six months.

Workflow

Built Coinbase's Incentives Engine and the growth loops, referrals, and onboarding primitives around it; now does growth engineering at OpenAI and teaches a course on engineering and shipping growth experiments.

Workflow

Runs Perplexity's growth on a 'growth product vs. growth marketing' split, with big quarterly bets aimed at a ~25% hit rate and activation instrumented around getting a new user to three queries in their first session.

Workflow

Built two marketing automations at Anthropic with Claude Code despite no prior coding: a Figma plugin that mass-generates ad-creative permutations (cutting creation from ~30 min to 30 sec) and an '/rsa' Google Ads workflow that pulls campaign data plus brand-tone skills and outputs an upload-ready CSV.

Workflow

Wires the Facebook Ads Marketing API into Claude Code so the agent bulk-uploads ads, kills losers, scales winners, spins up new ad sets, and reads performance to recommend creatives — no time in the Ads dashboard.

Workflow

Led Verkada's Clay-powered ABM machine: enriched email lists with LinkedIn URLs to lift ad match rates from ~50% to ~90%, scraped HOA directories into LinkedIn Ads audiences, auto-generated 600+ personalized Webflow landing pages, and ran a Clay-enriched direct-mail campaign driving $20-25M quarterly pipeline.

Workflow

Built Cursor's growth-experiment iteration loop: A/B testing the signed-out marketing funnel by reusing Statsig from the signed-in product, with the Vercel Flags SDK and Microfrontends, evaluated at low latency via Edge Config.

Workflow

Codified Superhuman's aha-moment method: regression analysis on retention vs. count of a core action to pick a 'line in the sand' ('50 emails cleared in 7 days'), paired with white-glove human onboarding that pushed 65%+ of new users to fully switch.

Workflow

Runs Ramp's AEO program as a portfolio of experiments: coin a proprietary citable metric ('Ramp Rate'), build programmatic data directories, and refresh pages for AI-search visibility, measuring lift per cohort (+78% LLM visibility on freshness, +24% on jobs-to-be-done content).

Workflow

Lays out his message-market-fit method: segment a market by what actually changes how you communicate (not demographics), then run rapid in-house cold-outbound tests to find the winning angle.

Workflow

Built Owner.com's AI sales stack: third-party restaurant data scraped into Snowflake plus Momentum-structured call transcripts feed prospect and ICP scoring, with AI filling fields (next steps, competitor mentions, ranked challenges) under human review before reps act.

Workflow

Ran the Notion PLG onboarding experiments she's personally credited with: a contextual 'Built with Notion' signup-button change that drove a 2.5x conversion lift, plus lifting the free-plan single-member cap to unlock collaboration.

Workflow

Breaks down Aikido’s viral April Fools “OnlyScans” rebrand: a produced reveal video plus an onlyscans.dev microsite she vibe-coded over a weekend in Lovable, as a repeatable “fun marketing” play.

Workflow

Built Braintrust's AI-search/AEO marketing engine: gap analysis on where competitors get cited in AI answers, tracking which URLs models reference, topic grouping, and a prioritized content pipeline, attributing real pipeline to AI search as its own channel.

Workflow

Builds thin-harness / fat-skills Claude Code agents for growth and content ops — including this directory's Apify-to-Clay-to-classifier pipeline, a weekly "content radar" that surfaces overlaps between internal signals and live trends, and an automated LinkedIn influencer-discovery engine.

Workflow

Created a 'power-law user acquisition' framework from her Meta/Pinterest/Descript work: identify a product's natural distribution advantage, match it to channels via B2B/B2C decision matrices, then concentrate on the 1-2 channels that drive 70%+ of growth.

Workflow

Created and led a hands-on 3-part 'build advanced AI agents' course: a real-time voice sales agent, a web-search research assistant, and a multi-agent user-research workflow that auto-generates personas, runs interviews, and synthesizes insights.

Workflow

Treats Vanta's GTM as a state machine: a one-click, Salesforce-triggered QBR-deck generator that pulls usage and funding signals into Google Slides (triggered 1,500+ times) and an AI-SDR pipeline in Clay that enriches, ICP-scores, routes to personas, and auto-queues email plus LinkedIn tasks.

Workflow

Built Oleve's growth experimentation machine: 6-7 simultaneous A/B tests on the flagship app (10+ daily across products) via SuperWall paywall tests and Mixpanel, plus reverse-engineered TikTok viral playbooks and AI agents for market research and creator sourcing.