GTM Engineer building AI-scored outbound systems on Clay, HubSpot and n8n
An anonymous operator open to contract work, based in Canada / USA (Eastern time, remote), working remote, available now for 20h/wk, working with SQL, Instantly, HeyReach, Nooks, Pipedrive. Rate: $75/hr USD. Vetted by AgentOps; name and contact details are shared once the operator accepts an intro.
- Rate
- $75/hr USD
- Availability
- available now (20h/wk)
- Region
- Canada / USA (Eastern time, remote)
- Working
- Remote
- Open to
- Contract
- Stack
- SQL, Instantly, HeyReach, Nooks, Pipedrive, Clay, n8n, Make, HubSpot, Claude Code, Python, Apollo, Salesforce
- H2A ratio
- 1:10 (humans to AI agents)
- Brings own agents
- Yes
Agents built
- ICP scoring and LinkedIn outbound pipeline (Clay, Apollo, HubSpot, HeyReach, webhooks): Scores every prospect 0-100 across 10 persona tables using firmographics, tech stack, hiring and growth signals pulled by an AI research agent, writes icp_score and lead_tier to HubSpot, and routes top tiers into LinkedIn campaigns. Connection and reply events webhook back into Clay so the CRM stays current with no manual exports. Outcome: 4,500+ prospects scored under a 5-credit cap per record · 10 campaign-ready persona lists · zero manual CSV syncing
- Per-contact cold email writer (Clay AI columns, Claude, HubSpot, Instantly): Turns CRM fit data (fit score, segment, vertical, named competitors) into two-sentence cold emails for each contact. Deterministic opener rotation keeps hundreds of rows from reading the same, it branches on missing data instead of inventing it, and cleanup columns strip AI tells before the copy lands in Instantly custom variables. Outcome: Hundreds of unique, human-sounding emails per campaign with no manual writing · every row passes an automated AI-tell QA pass before send
- Multichannel sequence system on AI-scored CRM data (HubSpot, Nooks, n8n): Takes an AI-scored company and buyer universe (fit score, segment, canonical buyer role) synced from HubSpot into Nooks and runs seven sequences on top: three baseline multichannel variants (LinkedIn, call, SMS, email), a gift dispatch and follow-up pair, and two reactivation plays driven by a loss-reason merge token instead of cloned sequences. Outcome: 4,800 raw companies filtered to 537 fit-70+ target accounts and 1,017 verified buyers · 7 sequences built and ready to enroll
- Franchise lead capture and follow-up automation (Pipedrive, Make): Routes paid social lead ads and website inquiries into a franchise-recruitment CRM, dedupes on email, creates a deal only when a meeting is booked, and runs step-counter follow-up automations through the connected mailbox, with compliant bulk sends through Campaigns. Outcome: Replaced a $99/mo connector with an $11/mo Make scenario · sized for ~100 inquiries/month before lead gen switched on
- Outbound ops skill library for Claude Code (Claude Code, custom skills, MCP (HubSpot, Clay, Pipedrive, Slack)): A set of Claude Code skills that encode my outbound playbook: campaign-type selection, Clay AI prompt writing, Clay-to-Instantly wiring checks, and a pre-send copy QA pass. Every client build runs through the same checks, with MCP connections into the CRM and sending tools so the agent works on live data. Outcome: Consistent builds across clients · client review rounds spend time on positioning instead of catching typos and AI tells
Proof of work
- Senior RevOps / GTM engineer at a B2B cloud infrastructure SaaS, owning CRM, enrichment and outbound automation. On the side I'm the activation layer for an AI data-engineering startup's clients: once their ICP and buyer scoring lands in HubSpot, I design and build the sequences, copy generation and automations that put it to work. Also contract directly with SMBs and agencies on Clay, HubSpot, Pipedrive, n8n and Make builds. Canada-based, Eastern time.