Founding Engineer

We're looking for someone whose default development loop is agentic: inspect the repo, plan, spin up worktrees, delegate/review sub-agents, write tests, run checks, ship PRs, and verify production behavior. You should already feel faster, sharper, and more ambitious because you build with coding agents every day.
Responsibilities:
Hardening the free AI-visibility audit into a dependable acquisition engine.
Improving audit latency, partial-result handling, queueing, backpressure, caching, and observability.
Building self-serve onboarding for brands, tenants, Slack, MCP, and API keys.
Creating curated MCP tools over the Brain API.
Turning raw SEO/AEO data into marketer-readable recommendations.
Building admin tools for audits, failures, spend, leads, and customer value.
Improving LLM grounding, report truthfulness, structured outputs, and evals.
Using Codex/Claude Code aggressively to ship faster while maintaining high verification standards.
Requirements:
Are genuinely Claude Code / Codex native.
Can take an ambiguous product problem and independently ship a tested production feature.
Know how to use agents for exploration, refactors, tests, docs, migrations, and code review.
Can manage multiple worktrees/branches without losing control.
Are strong in backend systems: queues, retries, DB models, API contracts, auth, PII, and observability.
Can build enough frontend to ship polished product flows.
Understand that AI-generated code must still be read, tested, and owned.
Have a good taste about what should be automated, what should be manual, and what shouldn't exist yet.
Care about truthfulness: no invented report claims, fake confidence, or ungrounded recommendations.
Prefer small, shippable PRs over massive speculative rewrites.

Experience
2-6 yrs