Research Collaboration Platform

A platform for managing datasets, experiment runs and publication drafts across several distributed research groups.

Client
Placeholder client
My role
Architect & Lead Developer
Year
2025
Category
Software Development
  • PHP
  • MySQL
  • Redis
  • Docker
  • S3

Every research group had its own shared drive and its own naming convention, and nobody could reproduce anyone else's results.

The core idea

Datasets are immutable and versioned. An experiment run records the exact dataset version, code commit and parameters it used, so a result can always be traced back to the thing that produced it.

Adoption

It succeeded because uploading through the platform was faster than the shared drive it replaced. Tools that are slower than the bad habit never win.

Outcome

Replaced a shared drive and a spreadsheet for around 120 researchers.

Related work

Logistics Forecasting Dashboard

Demand forecasting across a regional distribution network, with the forecast intervals shown as prominently as the point estimate.

Read more

Common questions

What kind of projects does Humza Imran take on?

Selected case studies are on this page. Work is usually applied AI, production software, or both, with a named outcome rather than a slide deck.

How can I work with Humza Imran?

Use the contact form at https://humzaimran.com/contact or email humzaimran505@gmail.com. Describe the problem, the timeline, and whether you need a system built, a team trained, or both.

Get in touch