Manufacturing · Data processing
Cutting the wait for factory inspection results
As part of the backend team, I helped reduce inspection-data processing time by approximately 50% and bring results from different machine types into one dataset.
I contributed to this project as part of an engineering team. This case study focuses on the areas I worked on and what I learned along the way.
What needed to improve
Inspection machines produced large amounts of data in different formats. Processing was too slow, and the results needed to be brought together before they could be analysed. I worked on both parts: reducing the processing time and making data from different machine types comparable.
What the system had to handle
- Data kept arriving from several types of inspection machine.
- Related results from different machines had to be matched.
- Slow processing meant a longer wait for results in the dashboard.
- The system needed to remain usable as new inspection data arrived.
My contributions
Find and address the processing delays
Slow processing meant people had to wait for inspection results. I helped identify where delays occurred and worked with the team on improvements, contributing to an approximately 50% reduction in processing time.
Give different machine results a common format
Faster processing addressed only part of the problem: results from different machine types also needed to be comparable. I worked on bringing them into a consistent format so they could be analysed together.
What went wrong
Testing needed to reflect real workloads
Early testing did not fully reflect the volume of data from the factory line. I helped address the performance problems that became apparent under that workload.
What I learned
I would test with realistic data volumes earlier. It helped me understand how differently a system can behave under everyday workloads compared with small test examples.