Solving VIRTUAL_ENV Directory Conflicts in UV Python Environments
Iāve been using uv to manage my Python environments. Itās a great tool, but I recently encountered an issue with the VIRTUAL_ENV environment variable. The Problem The issue arose when I was using uv to remove a package from what I thought was my activated virtual environment: (cbsim) ā cbsim git:(km/mle-1722/evaluation-harness) ā uv remove papermill warning: `VIRTUAL_ENV=/Users/khalil/Desktop/cbsim/.venv` does not match the project environment path `.venv` and will be ignored The error message revealed the root cause: my VIRTUAL_ENV environment variable was pointing to an absolute path that didnāt match uvās expected relative path. ...
The Schema Mismatch Problem with BigQuery Wildcard Tables
What are wildcard tables? Wildcard tables provide a concise way to query multiple tables that share a common naming pattern and schema. A wildcard table effectively represents aĀ UNIONĀ of all tables matching a specified wildcard expression (e.g.,Ā table_prefix*). When you query a wildcard table, BigQuery treats the matching tables as a single logical table. A special pseudo-column,Ā _TABLE_SUFFIX, is automatically included in the query results. This column contains the value matched by the wildcard character for each row, allowing you to identify the source table. ...

Monitoring Recommenders
Monitoring and observability systems track the performance and reliability of recommender systems in production. They report metrics computed from model inputs, outputs, and customer interactions. These metrics help organisations assess the health and effectiveness of their recommender systems. Failure to adopt a proper monitoring and observability system could negatively impact business operations affecting things like user experience and revenue. I believe there are three main components essential for monitoring a recommender system in production: Metrics, Logging, and Reporting. ...

AI Washing
The promise of generative AI is exaggerated in the short-term and underestimated in the long-term leading to an uncontrollable hype. Naturally, many choose to ride this hype with false advertisement to further their agenda ā whether it is to drive up stock prices or to win customers. BBC recently referred this false misrepresentation of the benefits of AI (to businesses) as āAI washingā: It is a problem that has quietly existed for a number of years, according to data from another tech investment firm, MMC Ventures. In a 2019 study it found that 40% of new tech firms that described themselves as āAI start-upsā in fact used virtually no AI at all. ...