Back to work

PERSONALIZATION PIPELINE

I built an AI personalization pipeline for outbound email — Python, web scraping, and LLM generation stitched into one flow. It researches each prospect, personalizes the angle, and drafts the message. The numbers moved.

AI AutomationOutbound EmailMeasurable Lift
+50–70%
Open rate lift
+30–40%
Click rate lift

// WHAT IT DOES

Most outbound is either generic-at-scale or personalized-but-slow. The pipeline collapses that tradeoff: it pulls public signal on each prospect, reasons about the most relevant hook, and drafts a message that reads like a human wrote it for one person.

Built in Python — scraping and enrichment feed an LLM generation step, with guardrails so the output stays on-voice and on-offer.

// THE PIPELINE

Research

Scrape & enrich public prospect signal.

Personalize

Pick the most relevant angle per prospect.

Draft

Generate an on-voice, ready-to-send message.

// TRY THE PERSONALIZER

Paste a public LinkedIn URL, pick a message type, and generate a draft.

Uses only your public LinkedIn profile info — nothing is stored.
Output
// generated draft will appear here