Secure Document Uploads with AI Summaries

I need a seasoned full-stack developer to roll out a production-ready document-upload feature that customers can use to submit files for review. After a file lands on the server the system must (1) store it safely, (2) restrict access so that every user sees only their own uploads, and (3) trigger an AI routine that returns a concise, plain-language summary of the document.

Scope of work
• Build a front-end widget or drop zone that handles large files and gives users real-time feedback.
• Create a hardened back-end API that streams the upload into encrypted storage, records metadata, and exposes retrieval endpoints.
• Implement authentication and per-user authorization; no file should ever be visible outside the owner’s account.
• Integrate an AI service (e.g., OpenAI, Claude, or a locally hosted model if you prefer) to generate and store a summary immediately after upload.
• Expose the summary through the same API so it can appear in the user interface alongside the original file.
• Add logging, basic audit trails, and automated tests to prove the security and reliability of the flow.
• Provide deployment scripts or a Docker setup so I can reproduce the environment in staging and production.

Acceptance criteria
1. A customer can upload PDF, DOCX, or image files up to the agreed size limit.
2. After upload, the customer sees a “processing” state and later a readable summary in plain English.
3. If the customer logs out and back in, they still see only their own files and summaries.
4. Attempting to fetch another user’s file or summary returns an appropriate 403/404 response.
5. A one-command deployment (or CI pipeline step) recreates the system on a fresh server.

If you already have experience with secure object storage (S3, GCS, Azure Blob, etc.), JWT/Session auth, and AI text generation, tell me which stack you’d reach for and roughly how long integration would take. I’m ready to start as soon as we agree on the approach and milestones.