PinnedAbuzar MahmoodProbabilistically determining the number of components for unsupervised learningAn excursion into Bayesian Nonparametrics for changepoint modellingOct 11, 2022Oct 11, 2022
PinnedAbuzar MahmoodWeeding out noise with MLUsing XGBoost for improving data preprocessing and semi-automating user data selectionJul 29, 2022Jul 29, 2022
Abuzar MahmoodLet AI do the work a.k.a. Rapid deployment using Streamlit appsIn many academic labs, despite our accelerating utilisation of computers, I have seen that a lot of important documentation still happens…Jun 24Jun 24
Abuzar MahmoodAutomating Data-Processing Pipeline Testing using Prefect 2.0For open-source contributions to work, someone needs to make sure that new additions to the code-base don’t break anything else. This is…Mar 15Mar 15
Abuzar MahmoodSupervised (a.k.a. Task-specific) dimensionality reductionWe often encounter data with more features/dimensions than we know what to do with. Many of them are correlated (even collinear), many of…Feb 19Feb 19
Abuzar MahmoodOverlySpecificGPT — Creating your custom LLM Chatbot with Retrieval Augmented GenerationThe rapid rise of Large Language Models (LLMs) has been both very exciting and slightly concerning. While it raises questions like “Will…Jan 12Jan 12
Abuzar MahmoodComparing Dimensionality Reduction Algorithms for User-Facing ProcessesIn a couple previous posts (Weeding out noise with ML, ML Experimentation and Model Registry with Neptune.ai), I mentioned how part of the…Mar 22, 2023Mar 22, 2023
Abuzar MahmoodThe need for data quality testing at the analysis level — A case study using neural data.When analyzing data, we usually have a battery of tests to make sure that the fundamental properties of the data are sound. This can look…Mar 7, 2023Mar 7, 2023
Abuzar MahmoodML Experimentation and Model Registry with Neptune.aiExperimentation forms an integral part of any complex process; the process of building machine learning models included. Building a…Aug 15, 2022Aug 15, 2022