GTM Engineer

You own the list. Not the application that displays it, not the outreach that goes out from it — **the list itself**, and the machinery that keeps it accurate, ranked, and fresh.

Why this role exists

We're a bootstrapped B2B services company. Our sharpest buying signal is a public event that tells us an account has just moved in-market — a ---------- nt when what we sell becomes urgent rather than interesting.

That signal is buried in third-party data feeds. Millions of records a week. Maybe two thousand matter to us.

Right now, finding those two thousand is guesswork wearing a lab coat. A query someone wrote once, a filter nobody has re-tuned since, and a number at the end that nobody can defend.

Your work is to make that number defensible, then make it bigger without making it worse.


What you'll own

1. The signal pipeline
Third-party data feeds ? filtered ? deduplicated ? scored ? landed in our database. You design the queries. You tune the filters. You own the credit budget and every coverage tradeoff inside it.

2. The account universe
Not just the inbound flow — the full stock of companies that could ever buy from us. Industry association directories, technographic sources, licensing registries, public company web sources. Feeds are rented. This asset is ours permanently.

3. The scoring model
Not every signal is equal. An account in our size band that just triggered our buying event is gold. An account that looks identical on paper but structurally cannot buy is noise, and noise is expensive because Sales spends real hours on it. You build the logic that separates them, and you revise it when Sales tells you it's wrong.

4. Enrichment
Company ? decision-maker ? verified contact. You own the waterfall, the match rates, and the cost per usable contact.

5. The truth
When someone asks "how many qualified signals did we get last week," you answer with a number you can walk backward to source. That sounds small. It's the entire role.



What you need to have done before

- Built and maintained a real B2B list pipeline — pulling from APIs, cleaning, deduplicating, enriching, shipping to a CRM or sequencer. Not a one-off extract. Something that ran daily and that people depended on.
- SQL, fluently. You can write the query that finds the duplicates.
- Python or equivalent for API work, pagination, rate limits, retries, and scheduled tasks.
- Worked with firmographic, technographic, or intent data providers — Apollo, Clay, ZoomInfo, People Data Labs, Bright Data, or similar. You know that a vendor's coverage claim and your actual result are two different numbers.
- Deduplication and entity resolution. Same company, four name variants, three domains. You've solved this at scale before.
- Cost discipline. You've operated inside a credit budget and made tradeoffs rather than just asking for a bigger plan.

What helps

- Exposure to B2B services, real estate, or professional services verticals
- Clay, Instantly, Supabase, or comparable tooling
- Technographic tools (BuiltWith, Wappalyzer)
- Comfort writing a short memo that says "our number is wrong, and here's why"

What this role is not

- Not a scraper-builder. We buy data where buying beats building. You'll need judgment about which is which, and the default answer is usually "buy."
- Not a full-stack developer. Engineering builds the product. You own the data moving through it.
- Not an SDR. You don't send the outreach. You make sure it reaches the right people.
- Not set-and-forget. The ICP moves. Vendors quietly change coverage. Filters rot. If you want to build something once and walk away, this will frustrate you.

---

How we'll measure it

- Qualified signals per week — and whether that number survives an audit
- Precision — Tier A accounts Sales agrees were worth a call
- Cost per qualified account
- Freshness — hours from signal appearing to account landing
- Duplicate rate — trending down, every month