Secure FastAPI AI Backend
Budget: $1,500 – $3,000 USD
We’re sourcing a backend developer to wrap our confidential AI engines with a clean, secure API layer. This is part of our internal AI assistant project for handling sensitive knowledge — all data stays on-premises or within our private cloud.
Scope / Responsibilities:
Build FastAPI-based endpoints that forward requests to internal AI models (LLaMA 3, Mixtral).
Implement JWT authentication with role-based access control, refresh tokens, and detailed request logging.
Integrate with MongoDB (or Postgres if preferred) for user data, metrics, and audit logs.
Create a small admin namespace for usage monitoring and basic maintenance tasks.
Deliver Docker-ready code so the stack can be deployed on private Kubernetes or other private infrastructure.
Requirements:
Python 3.x and FastAPI expertise.
Experience with JWT, role-based access, and API security best practices.
Familiarity with MongoDB or PostgreSQL, Docker, and unit testing.
Experience building internal or secure APIs is strongly preferred.
NDA / Confidentiality:
NDA required before any project specifics are shared.
Deliverables / Acceptance Criteria:
Endpoints return the expected payloads reliably under load.
Authentication tokens are correctly verified, refreshed, and revocable.
Logs and admin stats are accessible through the protected admin routes.
Repository builds and runs with one Docker command.
Scope / Responsibilities:
Build FastAPI-based endpoints that forward requests to internal AI models (LLaMA 3, Mixtral).
Implement JWT authentication with role-based access control, refresh tokens, and detailed request logging.
Integrate with MongoDB (or Postgres if preferred) for user data, metrics, and audit logs.
Create a small admin namespace for usage monitoring and basic maintenance tasks.
Deliver Docker-ready code so the stack can be deployed on private Kubernetes or other private infrastructure.
Requirements:
Python 3.x and FastAPI expertise.
Experience with JWT, role-based access, and API security best practices.
Familiarity with MongoDB or PostgreSQL, Docker, and unit testing.
Experience building internal or secure APIs is strongly preferred.
NDA / Confidentiality:
NDA required before any project specifics are shared.
Deliverables / Acceptance Criteria:
Endpoints return the expected payloads reliably under load.
Authentication tokens are correctly verified, refreshed, and revocable.
Logs and admin stats are accessible through the protected admin routes.
Repository builds and runs with one Docker command.