While building Vidyari—a digital platform I'm developing for study materials and courses—I hit a major architecture wall. Users needed to upload heavy PDFs and video lectures, but pushing massive files directly through my Node.js and Express backend was eating up server memory and causing timeouts.
I quickly realized the server shouldn't be the middleman for heavy lifting.
To solve the bottleneck, I completely re-architected the upload flow using AWS S3 Multipart Uploads. Here is how it works now:
1️⃣ The client requests a secure presigned URL from the backend.
2️⃣ The large file is sliced into smaller chunks on the frontend.
3️⃣ Those chunks are uploaded in parallel directly to the S3 bucket—completely bypassing my Node server.
The best part? If a user's Wi-Fi drops at 99%, the system only retries the failed chunk instead of starting the entire upload over from scratch.
Building Vidyari is teaching me that making something work locally is just step one.
Designing it to handle real-world conditions without crashing is the actual challenge.
For the backend devs out there: what's your go-to strategy for handling large media uploads efficiently? Would love to hear your insights! 👇
what to check out real word postmortems 👀
Top comments (0)