← All work
Team projectPaper under review

AI-Driven Auto Deployment Tool

GitHub URL in, running app out. It writes the Dockerfile and the CI pipeline, deploys to Kubernetes, and proposes a fix when the build fails.

FastAPI · React · Docker · GitHub Actions · Kubernetes · MongoDB

  • 7 language templatesSource: docs/paper-claim-verification.md
  • 3 LLM providers in fallbackSource: Groq, then OpenRouter, then Gemini; repo README
  • IEEE paper under reviewSource: Co-author; submission not yet decided

The problem

Getting a student project from a repository to a running service takes a Dockerfile, a registry, a pipeline and a cluster. Most people stop at the first one.

What was built

  • Detects the stack with a confidence score, then generates a Dockerfile from 7 language templates with an LLM fallback.Source: Repo README and docs/paper-claim-verification.md open
  • Generates a GitHub Actions pipeline that pushes to GHCR, deploys to Kubernetes on Minikube, and streams logs to the browser over SSE.Source: Repo README; pipeline templates in backend/
  • Turns a failed CI log into a one-click LLM patch, with provider fallback from Groq to OpenRouter to Gemini.Source: Repo README and docs/paper-claim-verification.md
  • I fixed the GHCR image-naming and entry-point path bugs that broke deploys from accounts that did not own the repo.Source: My commits of 25 July 2026
  • I audited the paper's claims against the code, found that "success" counted pods running and not apps responding, and redesigned the evaluation.Source: docs/paper-claim-verification.md in the paper worktree

What this does not show

  • This is a team project and my repository is a fork. A teammate wrote most of the commits.
  • My contribution is the deploy fixes, the claim audit and the evaluation redesign. I am a co-author on the paper.
  • The paper is under review. It has not been accepted.
  • Deploys target a local Minikube cluster, not a managed cloud cluster.

Receipts

ClaimWhere it comes from
7 language templatesdocs/paper-claim-verification.md
3 LLM providers in fallbackGroq, then OpenRouter, then Gemini; repo README
IEEE paper under reviewCo-author; submission not yet decided
Detects the stack with a confidence score, then generates a Dockerfile from 7 language templates with an LLM fallback.Repo README and docs/paper-claim-verification.md open
Generates a GitHub Actions pipeline that pushes to GHCR, deploys to Kubernetes on Minikube, and streams logs to the browser over SSE.Repo README; pipeline templates in backend/
Turns a failed CI log into a one-click LLM patch, with provider fallback from Groq to OpenRouter to Gemini.Repo README and docs/paper-claim-verification.md
I fixed the GHCR image-naming and entry-point path bugs that broke deploys from accounts that did not own the repo.My commits of 25 July 2026
I audited the paper's claims against the code, found that "success" counted pods running and not apps responding, and redesigned the evaluation.docs/paper-claim-verification.md in the paper worktree
NextReportGen: a project report from a short form