Founding AI Engineer & Architect at The Link Housing – Remote | Jobyssey

The Link Housing runs on a custom Claude-powered agent stack. Our 8-agent SAM.gov pipeline scans federal solicitations daily, extracts requirements from fillable PDFs, and generates response drafts. Our EA agent triages inbound email across several Gmail accounts, creates draft replies, and syncs to Monday.com. Orchestration runs on a dedicated mini-PC with cron-scheduled workflows, MCP servers, and integrations with SAM.gov, Google Places, and our CRM.

The next 12 months: scale it. More agents, better reliability, customer-facing AI tools (intake, tenant portal, property matching), and new verticals. You own the AI roadmap.

The team

We're a lean operating team of ~8 people handling property submissions, tenant placements, and landlord outreach. The founder architected the current AI stack and will be your direct technical counterpart. You'll be the first dedicated AI/engineering hire. No middle management, no tickets, no committee reviews — just ship.

Required

5+ years of software engineering experience, with at least 2-3 years building production LLM applications
Deep, hands-on experience with the Anthropic Claude API — you've shipped real systems with it, not just experimented
Demonstrated ability to design and own multi-agent systems end-to-end (orchestration, tool use, error handling, evals, observability)
Strong opinions about agent architecture, prompt engineering, and how to make LLM systems reliable in production
Comfortable being the most senior technical person on AI in the room, including with the founder
Strong Node.js skills, or Python with willingness to work in Node
Excellent written and verbal English — you'll be in customer conversations, partner meetings, and architectural discussions

Bonus

Experience with MCP (Model Context Protocol) servers
Background in real estate, insurance, or government contracting
Open-source contributions to AI/LLM tooling
Prompt evaluation or testing framework experience
PDF parsing pipelines (pdf-lib, pdftotext, pdf-parse)
Windows server / Task Scheduler familiarity

Mindset You think in systems, ship fast, and care about reliability. You're equally comfortable debugging a flaky API integration and designing a new agent architecture from scratch.

You're hungry. You read the Anthropic changelog the day it drops. You've tried MCP. You have opinions about prompt structure. AI isn't your job — it's the thing you'd be doing in your free time anyway.

You take ownership. If a customer-facing agent fails, you don't wait to be told — you find out, fix it, and write up what changed. You think like a founder. You don't just ship features; you think about what shipping this feature does for the business. You ask "what would 10x this?" and then go build it. You see the AI infrastructure as a competitive moat, not a tech stack.

Real ownership. First dedicated AI/engineering hire. You'll architect, ship, and own the AI roadmap end-to-end. Direct line to the founder, no middle management, no committee design reviews.

Competitive compensation. Annual performance reviews with clear paths for raises tied to impact.

Fully remote, work from anywhere. We don't care where you live as long as you can overlap a few hours with PST for live syncs.

Flexible schedule. Async-first culture. We care about what you ship, not when you're online.

Real business, real impact. We're profitable. Your work directly affects revenue, not a runway clock. The AI you build helps real families find homes after losing theirs.

Modern stack, modern tools. Claude API, MCP, Node.js, no legacy baggage. You'll work with current tech and have budget for tools you need.

Explore more jobs