Meeting Minutes
Date: Sunday, January 21, 2024
Participants: Clavon, Simon, Lucas, Evong
Agenda Items:
1. Review and discuss TDO feedback
2. Installation of Python packages/libraries
3. Data usage for model training
Key Discussion Points:
1. Review and discuss TDO feedback
– Clavon clarified that DTS is the infrastructure/application owner, while EO owns the TDO itself. DTS can’t provide specifics on the technical questions since they are infrastructure, but can coordinate with EO.
– Evong asked for more details on the investigation database – what data it contains and how it will be accessed. Lucas explained it was created by PwC to help investigations team automate processing of failed payment messages. It contains some customer information. The machine learning model will try to extract info from unstructured text fields to auto-populate form fields for the investigations team. There will be human validation after the ML predictions.
– Discussed information classification and filtering of customer data. Lucas clarified no data removal, just auto-populating of empty fields from unstructured data. Customer data is present in source messages.
– Evong summarized engagement checklist items that need to be evidenced throughout project lifecycle before close. Lucas acknowledged these steps need to be completed.
2. Installation of Python packages/libraries
– Lucas requested guidance on process to install Python libraries needed for project. Packages need to be scanned before installation.
– Evong will discuss with Sean on Monday to determine who will facilitate scanning and installation. Suggested submitting request to IT team to download packages from outside and bring into secure network segment for scanning.
– Lucas asked if pip install can be done to download packages. Evong confirmed request needs to go through IT given network restrictions.
– Follow-up meeting to be held on Tuesday with Sean’s team to discuss next steps.
3. Data usage for model training
– Lucas explained intention to move production data to development environment for model training.
– Evong stated this normally isn’t allowed and would need a policy exception. Suggested discussing with data owners and submitting exception form. Data protection assessment may be required as part of exception approval.
Decisions Made:
– Evong to discuss scanning and installation of Python packages with Sean on Monday.
– Follow-up meeting to be scheduled for Tuesday with Sean’s team to discuss package installation.
– Lucas to submit request to IT team to download needed Python packages from outside network.
– Policy exception required for moving production data to development environment. Lucas to engage with data owners.
Action Items:
– Evong to discuss package scanning with Sean and determine responsible team.
– Evong to look up policy exception form and send to Lucas.
– Lucas to submit request to IT for package download.
– Lucas to schedule Tuesday follow-up meeting with Sean’s team.
– Lucas to engage data owners on moving production data for policy exception.
