Mobile AI Bot Backend Development

I need a rock-solid backend that powers an AI bot dedicated to running automated processes inside my mobile application. The service must expose clean, well-documented APIs that my iOS and Android teams can call for task orchestration, status checks, and user-triggered actions.

Core capabilities I expect:
• Natural Language Processing so users can phrase commands in plain English.
• Machine Learning modules that continuously refine task execution rules from usage data.
• Real-time notifications delivered through Firebase / APNs to keep users informed the instant a process completes, fails, or needs input.

I am open on the final tech stack, though Python (FastAPI) or Node.js (NestJS) paired with a Postgres or MongoDB store would fit nicely with the rest of our setup. Whatever you choose, containerised deployment (Docker), automated testing, and CI/CD hooks to GitHub Actions are non-negotiable.

Deliverables:
1. Source code and Dockerfile for the backend service.
2. REST or GraphQL API documentation (Swagger or similar).
3. Trained NLP/ML models with clearly explained training pipeline.
4. Integration guide that lets my mobile team plug in and validate end-to-end flows within one sprint.

Acceptance is complete when the mobile app can trigger an automated process, receive a live progress push, and see the outcome reflected in under five seconds on average.