Published inBetter ProgrammingI Forked “Asyncpg” — And It Parses Database Records to Numpy 20x FasterPresenting the fork of asyncpg — asynchronous PostgreSQL client for Python — based on NumPy structured arraysJun 2, 20222Jun 2, 20222
Published inTowards Data ScienceMy Continuous Integration takes too much time. How do I fix it?Approaches to reducing the CI run times. Measure, optimize, leverage the hidden resources, and more.Mar 23, 20223Mar 23, 20223
Published inTowards Data ScienceHow we optimized PostgreSQL queries 100xSome of the tricks we used to speed up SELECT-s in PostgreSQL: LEFT JOIN with redundant conditions, VALUES, extended statistics, primary…Mar 8, 20223Mar 8, 20223
Published inBetter ProgrammingContinuous Performance Improvement of HTTP APIWe developed an analytics HTTP API in Python. This is how we discovered what to improve to serve requests fasterFeb 23, 20223Feb 23, 20223
Published inTowards Data ScienceHow we optimized Python API server code 100xSome of the tricks we used to speed up calls to our analytics API written in Python: played with asyncio, messed with SQLAlchemy, hacked…Feb 9, 20224Feb 9, 20224
Should you enlarge your PRs?The empirical laws of how long it takes to review and release a pull request depending on its size. We inferred them from 100,000 pull…Jun 10, 2021Jun 10, 2021
Published inTowards Data ScienceDeploying Kubeflow to a bare-metal GPU cluster from scratchMy experience of deploying Google’s Kubernetes ML toolkit on physical servers with multiple GPUs: tuning the host OS, installing k8s…Mar 18, 20212Mar 18, 20212
Published inTowards Data ScienceFuzzy matching people namesI’ve recently had to solve an interesting problem: given two unordered lists with real people names, match identities in between.Feb 25, 20211Feb 25, 20211
Published inTowards Data ScienceHacking Google Coral Edge TPU: motion blur and Lanczos resizeEdge TPU is a powerful device that is capable of general-purpose image filtering thanks to efficient 2D convolutions and TensorFlow.Nov 28, 2019Nov 28, 2019