qubis -- team : all iz well

The Qubis is an intelligent software-engineering system that utilizes Reinforcement
Learning (RL) to dynamically prioritize automated tests in a Continuous Integration (CI) pipeline. Instead of
running an entire test suite sequentially or arbitrarily upon a code change, the agent makes adaptive,
sequential decisions to surface failures as quickly as possible.

Modern software projects often contain thousands of automated tests. Traditional CI pipelines adopt a "run
everything" approach on every push. For a project with 8,000 tests, a full run might take 40-60 minutes,
even if a developer only modified a single file like payment/service.py .
This traditional approach introduces significant operational friction:
Long Feedback Loops: Developers wait extended periods to discover if a recent change introduced
a bug, slowing the development cycle.
High Compute Costs: Executing thousands of unrelated tests consumes unnecessary CPU time,
memory, and cloud CI minutes.
Wasted Computation: Tests completely isolated from the changed code have a near-zero probability
of detecting a relevant fault, making their immediate execution inefficient.

The ultimate success of the system is measured by tangible engineering improvements:
Time to First Failure (TTFF): The wall-clock time required to discover the initial fault.
Compute Saved: The percentage reduction in CPU time, memory, and cloud costs compared to
executing the full test suite.
Fault Detection Rate (FDR): The percentage of actual, historical failures successfully detected by the
prioritized subset. Note: Reducing CI compute is only viable if the FDR remains uncompromisingly
high.


----- Team : All iz well 
team member : Souptik Pal , sayak adak , proma roychowdhuri ,urjita paul, tiyasa kamle