Full Time
$2000
40
Apr 10, 2026
Prisma + Supabase + API + Worker/Deployment Debugging
I have an application that is about 80% complete. It was built primarily through Claude, and I’m now at the stage where the remaining issues are beyond what I can confidently troubleshoot on my own.
I am **not** looking for someone to rewrite the app from scratch.
I am looking for a **senior-level developer / technical lead** who is comfortable working in an **AI-assisted development workflow**, where **Claude continues generating most of the code** and you act as the architect, debugger, reviewer, and decision-maker.
I need someone who can:
- review an existing AI-generated codebase
- diagnose bugs and architectural issues
- troubleshoot worker-service, database, Prisma, schema, API, and deployment/container issues
- tell Claude exactly what to do next
- review Claude’s output and refine prompts/instructions as needed
- help me move the application from “mostly built” to stable and production-ready
Important ...
This is **not** a role for a junior developer.
This is also **not** a role for someone who is uncomfortable with AI-generated code.
The ideal person is someone who can look at an existing codebase, quickly understand what’s happening, identify root causes, and use AI tools effectively as a coding assistant rather than fight against them.
You should be strong in most of the following:
- TypeScript / JavaScript
- Prisma
- PostgreSQL / Supabase
- schema design and migrations
- API integrations
- worker/service debugging
- deployment containers
- debugging production/staging issues
- backend architecture
- async flows, queueing, and state issues
- reading and improving AI-generated code
Bonus if you have experience with:
- NestJS
- WebSockets / real-time systems
- rate limiting / retry logic
- observability / logs / metrics
- scaling service-based applications
- Claude Code or similar AI-assisted coding workflows
How we will work is:
- you diagnose and architect
- Claude writes most of the code
- you review and direct Claude
- you step in manually only when necessary to unblock or correct the implementation
In other words:
**Claude is the implementation engine; you are the senior engineer and technical lead.**
In the first phase, I want help with:
- identifying the highest-priority bugs/blockers
- stabilizing the worker/database/deployment issues
- clarifying what Claude should do next
- creating a reliable workflow so I can keep building without getting stuck
Ideal candidate:
- highly practical
- excellent at debugging
- comfortable leading AI-generated development
- able to translate vague technical problems into precise implementation steps
- not thrown off by partially messy or AI-written code
- strong enough to spot when the architecture itself is the issue
If you are a great fit, this can turn into a higher paying, long-term / ongoing role.
Please do **not** apply if you are mainly a prompt hobbyist or a junior developer using AI as a Vibecoder.
I need someone who can truly diagnose systems (without relying on AI), then direct the build intelligently.
-------------------------------------------------
Technical brief linked below. Review it BEFORE applying.
-------------------------------------------------
TO APPLY, please submit answers:
1. What experience do you have debugging Prisma, schema, database, or deployment issues in existing apps?
2. Have you worked in a Claude Code or Abacus / AI-assisted workflow before? If so, how do you use AI effectively without letting it create more problems?
3. How would you approach a codebase that is 90% complete but has worker-service and database bugs?
4. Are you comfortable acting as the senior technical lead while AI handles much of the implementation/writing?
5. What is your English speaking, reading, and writing level, and how many hours a day do you have overlap with the US time zones?