NLP Text-Generation AI Service

I’m setting up an AI-powered service whose core skill is natural-language text generation on par with ChatGPT. What I need is the complete pipeline: model selection or fine-tuning, an efficient inference API, and a simple front-end that lets end-users submit a prompt and receive an articulate, context-aware response within seconds.

Here’s how I picture the workflow: you help me choose the right model family (OpenAI GPT-4, an open-source Llama variant, or any alternative you recommend), tune it for my domain, wrap it in a lightweight REST or GraphQL API, and deploy it to a scalable environment—AWS, GCP, or Azure are all acceptable as long as latency stays low and usage costs are transparent.

Deliverables
• Fine-tuned text-generation model with documented hyper-parameters
• Production-ready API (authentication, usage metering, error handling)
• Minimal web demo or Postman collection for quick testing
• Deployment scripts/infra as code so I can reproduce the stack
• Brief hand-off guide covering model updates and monitoring

Acceptance criteria
1. Prompts under 500 tokens return coherent answers < 3 s 90 % of the time.
2. Output must respect the temperature, max-tokens, and stop-sequence settings exposed in the API.
3. Codebase is clear, commented, and runs via a single command after installing dependencies.

If this aligns with your expertise in NLP, transformers, and cloud deployment, I’m ready to kick off as soon as you share a concise plan and timeline.