This is your work, valued

Melbourne, Australia

pgalko

Expert
@pgalko

If knowledge is power, knowing what we don’t know is wisdom

BambooAI. A Python library powered by Language Models (LLMs) for conversational data discovery and analysis.

782

athlete_data_warehouse. A tool for bulk download, formatting and SQL DB storage of exercise, training, nutrition, wellness and health data from various trackers.

31

Next-Number-Prediction. This is an experiment evaluating the performance of GPT-3.5-Turbo compared to a simple persistence model in predicting the number that follows a sequence of numbers..

27

query_fit_data_using_nl. This experimental project harnesses the OpenAI API to enable coaches and athletes to query data from FIT files stored in an SQL database using natural language, while also integrating relevant research papers for added context. The goal is to create a more accessible and user-friendly interface for analyzing and understanding training data.

21

Aerobic_Threshold_Estimation_from_DFA_Alpha1. A Python script that estimates aerobic threshold HR from activity HRV data. The script iterates through a list of activities and computes DFA Alpha 1 for each, then computes HR from alpha 1 (0.75) and plots the results. Takes a while to run.

7

ml_running_efficiency_predict. Jupyter Notebook - Optimizing Running Performance and Efficiency with Neural Network Models

2

ml_rest_hr_prediction_transformer. Jupiter Notebook - Predicting Resting HR timeseries using transformer model

1

Next-Number-Prediction-2. This is an experiment evaluating the performance of GPT-3.5-Turbo in predicting whether the next number in a sequence will be higher, lower, or equal to the last number.

1

delv-e. CLI tool for autonomous data exploration and analysis - iteratively generates hypotheses, writes and runs analysis code, and adapts based on findings. Works with Anthropic, OpenAI, OpenRouter, and Ollama models.

1