Revenue engineering
Target lists chosen by fit, enrichment that actually resolves, sequences that stop when someone replies, and a CRM you can trust at the end of the month.
- Clay
- HubSpot
- Mailgun
- Deliverability
Alexandros Nomikos
Aadd alex.jpg
AI
Software
Automation
Aadd alex.jpg
GTM
Sales
Marketing
← move across →
One person. No handoff.
Proof, not claims
Same motion I'd run for you: build the list, enrich it, sequence it, keep the CRM honest, report the numbers.
Clay for enrichment, a sequencer I wrote myself (Mailgun + HubSpot sync, send jitter, auto-stop on reply), and a kanban to track it. The first batch showed zero opens and I assumed deliverability. The logs proved me wrong: every message was accepted and inboxed. Plain text cannot carry a tracking pixel, so open rate was never measurable — I replaced it with bounce rate, which is the number that can actually hurt you. Four of the seven conversations came inbound, from writing about the build in public.
How the two sides meet
The gap between "we need pipeline" and "we shipped the thing" is where weeks disappear. I work on both sides of it.
Target lists chosen by fit, enrichment that actually resolves, sequences that stop when someone replies, and a CRM you can trust at the end of the month.
When the tool doesn't exist, I write it. Full products with Claude Code and Cursor — booking flows, admin panels, AI agents that draft and wait for a human's yes.
Economics at AUEB. Pricing, margin and payback modelled before the build, not after. A degree is only worth something if it changes a decision.
01Case study
I didn't just build the website. I helped build the system around the business.
A working spa on Santorini, run by real owners, with real bookings on the line. I didn't just ship the booking software — I built the customer-facing loop around it: the site that sells the appointment, the systems that bring people to it, and now an AI agent that handles the reviews that bring the next customer.
Website → Bookings → QR funnel → Google → Instagram → Ads → Reviews → AI agent
In production · real bookings, real revenue
More work
What changes when the outbound pipeline behind my job search has to run at ten times the volume for a B2B SaaS company, and the one part I would refuse to change. The decision that reverses, what breaks first, and why five of eight components survive untouched.
Read the paper →
Rather than pay for a sequencer, I built one: Mailgun delivery, HubSpot sync, send jitter, automatic stop on reply, bounce handling, pipeline kanban. Deployed on Vercel in a weekend. Later added per-timezone sending, so each email lands in the recipient's morning rather than mine.
Sends in the recipient's local morning · bounce rate 0%
Paste a domain, get a structured prep pack: pain points, discovery questions, objections with responses, demo strategy, follow-up email, CRM notes. Built because I needed it for my own calls.
Target: 30 minutes of prep down to 2
A financial literacy app for parents and children: kids complete tasks, earn rewards, and spend them at real merchants through actual discounts. Third overall across the whole field.
1st in challenge · 3rd overall
Villas, transfers and island services — briefed, built and delivered directly with the owners. Where I learned that a website is a sales asset or it is nothing.
5 delivered · direct client work
My story
I grew up on Santorini, where every business has a season. That taught me early that businesses don't have time for software that doesn't work.
Read my story →Now
This page changes. If it looks stale, tell me.
Stack