Senior Claude Code Advisor (Equity Only, 3-5hrs week)

I'm Ivan Kovpak, founder/CEO of Unstuck Engine. We're building the infrastructure layer for B2B GTM — real-time signal processing, ICP/Persona scoring, waterfall enrichment, and audience orchestration. Revenue generating, raising a pre-seed round, moving extremely fast.

Here's what's different about us:

Our entire engineering team builds with Claude Code 10-12 hours/day. We have 17 detailed product specs. We ship features in hours, not sprints. Claude Code writes 80% of our code.

And that's exactly why we need you.

The problem AI can't solve:

Claude Code is incredible at writing code from specs. It's terrible at knowing what it doesn't know. It can't tell us when our Supabase Edge Function will break under 100 concurrent requests. It can't spot that our RLS policy has a bypass we haven't thought of. It can't warn us that our credit ledger has a race condition that will lose us money at scale.

We need a senior engineer who has built and broken systems at scale — someone who looks at our code and immediately sees the 3 things that will blow up in production.

What you'll actually do:

• Code review our PRs (3-5 per week). Not line-by-line nitpicking — architectural review. "This approach won't scale because X." "This has a security hole because Y." "You're going to hit Z edge case at 1000 customers."
• Answer the questions Claude Code can't. We only escalate to you after Claude Code has failed to solve a problem after 3 attempts. You get the hard problems, not the easy ones.
• Monthly architecture review (1 hour). Look at what we've built, tell us what's wrong, what will break next, what we should refactor before it becomes technical debt.
• On-call for critical decisions. "Should we use Postgres advisory locks or Redis for this?" "Is this RLS policy secure?" "Will this Edge Function timeout under load?" Quick async questions, quick async answers.

Our tech stack:

Supabase (Postgres + Edge Functions + Auth + Realtime), Next.js, n8n, Stripe (usage-based credits), CopilotKit + Claude API, Cloudflare Workers, PostHog, Upstash Redis, TypeScript

What you'll review:

• Supabase schema design and RLS policies
• Edge Function architecture (our entire external API)
• Enrichment waterfall engine (multi-provider, per-customer ranking)
• Credit/billing system (Stripe + internal ledger, fractional credits, concurrency)
• n8n workflow modifications (ICP/Persona scoring pipeline)
• MCP server distribution (Cloudflare Workers)
• Security: API key auth, tenant isolation, webhook signatures

Time commitment: 3-5 hours/week. Fully async. No standups, no Slack channels to monitor, no meetings except one monthly architecture review. You respond to PRs and questions when it fits your schedule.

Compensation: Advisor equity (vested). No cash. We pay for nothing except your brain.

Why this is worth your time (honestly):

You keep your $200K+ job. You spend 3-5 hours/week reviewing interesting code at a high-velocity AI startup. You get:

• "Technical Advisor at funded AI startup" on your LinkedIn. This signals you operate above your day job's level.
• Real 0-to-1 startup experience without the financial risk. When you're ready to make the leap to CTO or founder, you'll have done it before.
• Equity in a revenue-generating company in the hottest B2B category.
• Pure engineering problems. No Jira, no PMs, no ceremonies. Just "here's the code, what's wrong with it?"

This is for you if:

• You've built production systems that serve thousands of customers
• You've been the person who finds the bugs that junior engineers miss
• You can review a PR and spot the race condition in 5 minutes
• You've scaled Postgres/Supabase beyond hobby projects
• You want startup exposure without quitting your job

This is NOT for you if:

• You want to write code (we have Claude Code for that)
• You need more than 5 hours/week of engagement
• You've never operated a production system at scale
• You think code review means checking variable names

To apply: Pick any one of our specs at github.com/ivan-unstuck/unstuck/tree/main/docs/specs — read it and send me 3 things you'd change or flag as risks. That's it. If your feedback is sharp, we'll talk.
https://app.dover.com/apply/Unstuck%20Engine/994f74d4-f666-4bde-a970-0c6b82de3718?rs=72237129