Full-Stack Node.js/Next.js Developer — AI Video Generation Platform (Agentic AI experience required)

Overview We're looking for a full-stack developer to join an existing production web application that turns a business brief into a suite of hyper personalised marketing content, including PowerPoint decks, Word documents and customised video content. This is a live, actively-used tool, built and maintained entirely through agentic AI development (Claude Code) rather than traditional hand-coding alone. We need someone who can work the same way — not just read and edit code, but direct an AI coding agent to extend the system correctly, safely, and consistently with the existing architecture. You will work with the existing creative and technical team to build and maintain this application. This is a 3-month contract with strong potential to extend and potentially transition into a full time role, based on fit and ongoing scope. What you'll work on Extending the AI-driven script and shot-list generation pipeline (LLM prompt engineering, structured JSON outputs, a knowledge base of visual/brand guidelines) Image and video generation integrations (multiple interchangeable providers) The FFmpeg-based video assembly pipeline (concatenation, cross-dissolves, multi-track audio mixing, loudness normalization, on-screen text overlays) Admin tooling for reviewing, regenerating, and approving generated assets at every stage Azure infrastructure and deployment Bug fixes and reliability work on a system that runs long, resource-intensive background jobs — including debugging issues that only show up in production Tech stack Frontend Next.js 14 / React 18 Microsoft Authentication Library (MSAL) for Entra ID (Azure AD) sign-in Backend Node.js / Express (REST API, Server-Sent Events for long-running job progress) JWT-based auth, role-based access control FFmpeg (via ffmpeg-static/ffprobe-static) for all video/audio processing AI & generative media integrations Anthropic Claude (via Azure AI Foundry) — script and prompt generation OpenAI GPT Image models (via Azure) — image generation Kling AI — image-to-video generation Higgsfield AI — alternative image-to-video provider (Seedance model) ElevenLabs — text-to-speech voiceover, with word-level timing data Document & file generation PowerPoint (pptxgenjs), Word (docx, mammoth), PDF (pdf-lib), Excel (xlsx-js-style) Infrastructure Docker, deployed to Azure Container Apps (separate API and web containers) Azure Blob Storage + Azure Files (persistent storage) Azure AI Search, Microsoft Graph API, SharePoint integration GitHub Actions CI/CD (build via Azure Container Registry, auto-deploy on push) npm workspaces monorepo (separate api and web apps) Data File-based JSON store, synced to Blob Storage (no traditional database) Requirements Must have hands-on experience with agentic/AI-assisted development (e.g. Claude Code, Cursor, Antigravity, GitHub Copilot Workspace, or similar) — this is how the codebase is built and maintained, and you'll be expected to work the same way, not just review AI-written code after the fact Strong Node.js / Express backend experience Strong React / Next.js frontend experience Comfortable working with third-party AI/ML APIs (LLMs, image/video/voice generation) — reading provider docs critically, since they're frequently inaccurate and need verification against the real API Experience with FFmpeg or video/audio processing pipelines a strong plus Experience with Azure (Container Apps, Blob Storage, Entra ID) a strong plus Comfortable working in a live production codebase with real users Clear written communication — this is a remote engagement and most collaboration happens async Contract details Type: Fixed-length contract, 1 month, with strong possibility of extension based on fit and ongoing needs Hours: 40 hrs/week Rate: Start date: ASAP Timezone overlap needed: Ireland or UK preferred How to apply Please include in your proposal: Examples of production work built with an AI coding agent (Claude Code, Cursor, etc.) — describe your actual workflow, not just the finished output Relevant experience with Node.js/Next.js and any of the integrations listed above (FFmpeg, Azure, LLM APIs) Your availability, rate and location