VibeCodeCareers 
This is a fully remote, mid-level role at Campfire, an AI-native ERP for financial operations. The position involves taking AI-assisted code from the CEO and bringing it to production quality, ensuring it aligns with customer needs and operational standards.
Skills / Requirements
- Accounting Correctness
- AI-coding Tools
- Celery
- Code Review
- Django
- Error Handling
- Financial Integrations
- LLM Call Management
- Migrations
- PostgreSQL
- Production-grade Code
- Python
- React
- Sentry
- Startup Experience
- TypeScript
- Vibe Coding
Why Apply
This role is ideal for someone who thrives on AI coding tools, as you'll work with AI-generated code directly from the CEO. You'll need to understand AI's failure modes and improve code quality from AI outputs, making it a perfect fit for vibe coding enthusiasts.
What You'll Be Doing
You'll triage pull requests from the CEO, determining their intent and making them production-ready. This involves adding types, tests, fixing queries, and ensuring accounting correctness. You'll also improve AI-generated code quality and maintain reliability standards.
Working in San Francisco, United States
- San Francisco is a major tech hub, offering abundant networking opportunities with other tech professionals.
- The city is known for its vibrant startup ecosystem, making it an exciting place for tech innovation.
- Public transport options like BART and Muni make commuting within San Francisco manageable.
- San Francisco's diverse culture and numerous tech meetups provide a rich environment for professional growth.
Pay and Career Growth
The role offers the chance to work closely with the CEO, providing unique insights into high-level decision-making and customer-driven development. The company values low ego and high candor, fostering a culture of open communication and continuous improvement.
Benefits and Perks
- Close to CEO
- competitive salary
- Customer Priority Focus
- Equity
- Remote Work
Is This Role Right for You?
Good fit if you...
- You have over 4 years of experience shipping production Python code.
- You can handle AI-generated code and know how to refine it for production.
- You are comfortable pushing back on high-level code and offering constructive solutions.
May not be for you if...
- You lack experience with Python, Django, or Postgres.
- You prefer structured, well-defined tasks over ambiguous, fast-paced work.
- You are uncomfortable challenging code from senior leadership.
Original Job Description
About Campfire
Campfire is an AI-native ERP for financial operations. Accounting teams and fractional CFO firms at fast-growing companies use it to run their close, revenue recognition, AP/AR, consolidations and reporting. Ember, our AI layer, works alongside accountants on the ledger itself.
The Role
Our CEO ships code. With AI coding tools, they turn customer conversations into working prototypes and open PRs directly against our codebase. The ideas are good and the instincts come straight from customers. The code is vibe coded.
Your job is to take those PRs to production.
You’ll be the engineer who sits between the CEO’s editor and main. You’ll figure out what each PR is actually trying to do, then decide what to keep, rewrite or drop. You’ll get it through review, tests, migrations and deploy without losing the idea that made it worth building. You’ll turn a customer demo that half works into a feature that is correct on a real general ledger with real money in it.
This is a real engineering role with an unusual input stream. You’ll work closer to the CEO and to customer priorities than anyone else on the team.
What You’ll Do
Triage every CEO PR. Read the diff and work out the intent. Decide which it is: ship with fixes, rebuild properly, split into smaller PRs, or push back.
Make it production-grade. Add types, tests and error handling. Fix N+1 queries. Make sure the code follows existing patterns instead of inventing new ones. Generate migrations properly and make them safe to deploy.
Protect the ledger. Nothing posts into a closed period. Every model attribute the code touches actually exists. Approval and audit flows (draft queues, permissions) are respected. Integrations hit sandbox endpoints outside production. Accounting correctness is not negotiable, however good the demo looked.
Keep AI features honest. Every LLM call the CEO wires up gets token tracking and spend-limit enforcement before it ships.
Guard reliability. Errors route through our Sentry conventions, failures report once, and customer-fixable problems surface to customers instead of paging engineers.
Close the loop. When the CEO keeps making the same mistake, fix the root cause. That might mean better scaffolding, clearer conventions in the repo, or tooling that makes the AI generate better code from the start.
Say no well. Sometimes the right answer is “this shouldn’t ship.” You’ll explain why clearly and offer a better path to the same customer outcome.
What We’re Looking For
4+ years of shipping production Python. Django, Postgres and Celery strongly preferred; TypeScript/React a plus.
Excellent code review instincts. You can read a 15,000-line AI-generated diff and quickly find the three things that will break in production.
Taste and judgment about when to patch, when to rewrite and when to delete.
Low ego and high candor. You’ll push back on the CEO’s code routinely and need to be comfortable doing it.
The ability to reconstruct intent from incomplete specs, and to ask the one question that resolves the ambiguity.
Speed. The CEO opens PRs fast, and the queue shouldn’t become a graveyard.
Nice to Have
Accounting fluency: double-entry, accruals, ASC 606, month-end close.
Deep experience with AI coding tools, including knowing their failure modes well enough to prevent them upstream.
Experience with financial integrations (Plaid, Ramp, Brex, Stripe, payroll providers).
Early-stage startup experience.
