Rename tool to Inference Runtime

[David-Uka]

Summary

• rename the public tool surface from AI Inference Platform to Inference Runtime
• update Kubernetes deployment and service names to inference-api and inference-worker variants
• rename Docker image references, Prometheus job name, and Grafana dashboard identifiers
• align the GitHub repository name and description with the new tool name

Testing

• python3 -m py_compile api/app.py api/metrics.py api/model.py api/queue.py api/worker.py
• ruby -e 'require "yaml"; Dir["k8s/**/*.yaml"].sort.each { |f| YAML.load_stream(File.read(f)); puts "ok #{f}" }'

[gemini-code-assist]

Code Review

This pull request rebrands the project from "AI Inference Platform" to "Inference Runtime" across the documentation, FastAPI configuration, Docker image names, and Kubernetes manifests. The renaming covers deployments, services, horizontal pod autoscalers, and monitoring configurations. One piece of feedback suggests that for complete consistency, the underlying Prometheus metric names should also be updated to use the new prefix, which would require changes in the API metrics definitions and the corresponding Grafana dashboard queries.

[gemini-code-assist]

While the dashboard title and UID have been updated, the underlying metric queries still use the old ai_api_ prefix (e.g., ai_api_http_requests_total, ai_api_inference_latency_seconds_bucket). For consistency with the renaming of the service to inference-api, consider updating these metric names to use a prefix like inference_api_. This would likely require changes in api/metrics.py to rename the metrics at the source, and then updating the queries in this dashboard to match.