We’re a recruitment firm hiring on behalf of a client who needs a hands-on builder to take a custom CRM from prototype to production.
If you love iterating fast, fixing gnarly bugs, and shipping clean, reliable features, this one’s for you.
What You’ll Do:
Own the end-to-end build of a lightweight CRM, stabilising an existing Replit prototype and taking it to production.
Design clean, extensible data models and implement core features (pipelines, notes, tasks, email/activity logging, integrations).
Use AI coding tools (Cursor, Lovable) with strong prompt engineering to move faster without sacrificing quality.
Integrate OpenAI APIs, n8n automations, webhooks, and third-party SaaS platforms.
Squash bugs, add tests, and improve DX/CI for maintainability.
Plan the migration from Replit/Lovable to enterprise-grade hosting (AWS/GCP/Azure/Vercel) or to established CRM platforms.
Must-Have Skills:
Full-stack fundamentals:
Frontend, backend, and data fluency.
Replit expertise: env setup, debugging, deploys.
Cursor & Lovable experience (AI IDEs).
Strong prompt engineering & LLM integration know-how.
APIs: REST/JSON, OAuth, webhooks.
OpenAI API: structured outputs, tool calling, rate-limit handling.
n8n: workflow design, retries, and error handling.
Debugging mindset turning “constant error” into “all green.”
Enterprise Readiness & Portability:
Cloud deployment: Migrate services from Replit to AWS (ECS/Fargate/Lambda), GCP (Cloud Run), Azure (App Service), or Vercel/.
Containers & CI/CD: Docker, GitHub Actions, environment promotion (dev → staging → prod), rollbacks.
Data migration: Schema versioning, seeding, backup/restore, bulk APIs.
Identity & Security: SSO/SAML/OAuth (Okta, Azure AD, Google), RBAC, audit logging, SOC2/GDPR hygiene.
Observability: Structured logging, metrics, Sentry, uptime checks.
iPaaS/ETL: n8n (primary), Zapier/Make/Workato familiarity helpful.
CRM Ecosystems: Salesforce, HubSpot, Pipedrive, Zoho (object models & APIs).
Infra as Code: Terraform basics, secrets management.
Data stack: PostgreSQL/SQLite now; ability to migrate to managed or analytical stores (Postgres/Snowflake/BigQuery).
Prompt Engineering — What “Great” Looks Like:
Prompt architecture: System/developer/user roles, templates, guardrails.
Reasoning patterns: ReAct, Chain-of-Thought, Tool Calling, RAG.
Structured outputs: JSON Schema, Zod, Pydantic; robust parsing.
Context management: Embeddings, vector search, caching, budgeting.
Evaluation: Golden sets, promptfoo/Ragas, A/Bs, cost/latency trade-offs.
Reliability: Jailbreak resistance, moderation, deterministic retries.
LLM tooling: LangChain, LlamaIndex, DSPy, Guidance, Instructor, Langfuse/Helicone observability.
Automation: Connecting LLM logic to n8n flows with resilient error handling.
The Vibe:
Builder energy: You ship, fix, and improve — fast.
Pragmatic quality: Clean, readable, well-tested code.
Product sense: You make smart, simple UX calls without over-engineering.
Low-drama, high-ownership collaboration style.
Logistics:
Engagement: Contract/freelance (with potential extension).
Location: Remote-friendly (AEDT overlap preferred).
Start: ASAP.
How to Apply:
Email:
Subject: CRM Vibe Coder - [Your Name]
Include:
Links: GitHub / Replit / portfolio + any n8n workflows you’re proud of.
A short note covering:
a) A time you stabilised a flaky prototype.
b) Your go-to stack for a small CRM.
c) How you use Cursor/Lovable + prompt engineering daily.
d) How you’d migrate a Replit app to AWS/GCP/Azure or a CRM like Salesforce/HubSpot.