Andrew J.

I'm a software engineer who does cold email for a living. Seven years writing code, the last few building outbound systems for B2B companies. Sourcing, scrapers, enrichment, deliverability, copy, and the campaigns themselves. About a dozen running right now. Most outbound freelancers can only do what their tools let them click. When the data isn't in Apollo, they're stuck. When the personalization needs to read a careers page or a permit registry, they're stuck. I just write it. ➡️ 7+ years as a software engineer, Master's in Computer Science ➡️ Around a dozen B2B outbound clients on month to month retainers ➡️ Founder of Mindaptive, an outbound and AI automation team 𝐖𝐇𝐄𝐑𝐄 𝐓𝐇𝐄 𝐄𝐍𝐆𝐈𝐍𝐄𝐄𝐑𝐈𝐍𝐆 𝐒𝐇𝐎𝐖𝐒 𝐔𝐏 🐍 𝐃𝐚𝐭𝐚 𝐩𝐚𝐬𝐭 𝐰𝐡𝐚𝐭 𝐀𝐩𝐨𝐥𝐥𝐨 𝐠𝐢𝐯𝐞𝐬 𝐲𝐨𝐮 I write Python scrapers for lists nobody sells. Association directories, job boards, review sites, government registries, niche marketplaces. Then I dedupe against your CRM so you're not cold emailing a current customer. That part gets skipped constantly and it's embarrassing when it goes wrong. ⚙️ 𝐌𝐲 𝐨𝐰𝐧 𝐞𝐧𝐫𝐢𝐜𝐡𝐦𝐞𝐧𝐭 𝐥𝐚𝐲𝐞𝐫 𝐢𝐧𝐬𝐭𝐞𝐚𝐝 𝐨𝐟 𝐫𝐞𝐧𝐭𝐢𝐧𝐠 𝐂𝐥𝐚𝐲 Clay is a good product. It's also around $500 a month on the mid plan before you buy a single row of data, and since the March pricing change you pay platform fees even on your own API keys. I run the same waterfall logic in n8n against the data APIs directly. You pay per call. No seat, no meter. When a workflow needs something no-code can't do, I write it. ✅ 𝐕𝐞𝐫𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧 𝐈 𝐜𝐨𝐧𝐭𝐫𝐨𝐥 Lists go through MillionVerifier and ZeroBounce, plus a checker I built in Claude Code that handles SMTP and catch-all logic before I pay for anything. Bad data is the number one reason sending domains die. People keep buying it anyway. 📬 𝐃𝐞𝐥𝐢𝐯𝐞𝐫𝐚𝐛𝐢𝐥𝐢𝐭𝐲 𝐚𝐬 𝐚𝐧 𝐢𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦 SPF, DKIM, DMARC, custom tracking domains, subdomain strategy, inbox rotation, warmup ramps, sending caps. I've picked apart setups where the copy was fine and the DNS was quietly killing every campaign. If yours is burning, let me read the records before you buy more leads. 🤖 𝐏𝐞𝐫𝐬𝐨𝐧𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐭𝐡𝐚𝐭 𝐫𝐮𝐧𝐬 𝐨𝐧 𝐫𝐞𝐚𝐥 𝐬𝐢𝐠𝐧𝐚𝐥𝐬 n8n workflows pull a prospect's site, LinkedIn activity, recent hires, funding, tech stack changes. That goes to Claude or GPT with tight prompts and a validation step, so nothing hallucinated reaches a prospect. Everyone skips the validation step. It's what keeps you from sending 400 people a compliment about a product they don't sell. 🎯 𝐖𝐇𝐀𝐓 𝐈 𝐑𝐔𝐍 𝐃𝐀𝐘 𝐓𝐎 𝐃𝐀𝐘 ➡️ ICP definition, offer angles, sequences, subject line tests ➡️ Follow ups, reply handling, weekly reporting ➡️ LinkedIn through HeyReach, synced with email so nobody gets hit twice in one morning ➡️ After the reply: CRM routing, lead scoring, booking links, no show follow up ⚠️ One thing I'll be blunt about. If someone is selling you on open rates, be careful. Open tracking broke when Apple started pre-loading images, and the pixel itself hurts deliverability. I report reply rate, positive reply rate, and meetings booked. 🧰 𝐌𝐘 𝐒𝐓𝐀𝐂𝐊 Sending: Smartlead, Instantly Data: Apollo, Sales Navigator, custom scrapers LinkedIn: HeyReach, Aimfox, Prosp, Dripify Automation: n8n, Make Verification: MillionVerifier, ZeroBounce, internal tool AI: Anthropic API, OpenAI API, LangChain Code: Python, TypeScript, Next.js, FastAPI, PostgreSQL, Supabase, Docker Tooling: Claude Code, Cursor 🎓 𝐁𝐀𝐂𝐊𝐆𝐑𝐎𝐔𝐍𝐃 Before outbound I spent years on enterprise healthcare software, the kind used by 90% of the hospitals in a country. Strict data handling, systems that aren't allowed to go down. Master's in Computer Science, top of my class. I run Mindaptive now and I still write code every day. 📍 Tell me your offer and who you're trying to reach. I'll tell you whether cold email is even the right channel for it, and if it isn't, I'd rather say so than take the money.