Financial Data Series Pattern Recognition Machine Learning Python Program

I'm looking for a Machine Learning and Python expert to build a system that can train and recognize patterns in stock market data. Key components of the project: 1. Python program to build machine learning model which will be further used to recognize patterns in data (program 1) 2. Python program to use the model built by the previous to recognize patterns in out-of-sample data (program 2) High level description - the programs must be implemented in Python, and be compatible with Python 3.12 or higher - number of patters to be recognized can be between 1 and a X, It is understood and acceptable that the more patterns are provided in the training data, the more time will be required to train the model. There will be no limit as to how many different patterns can be recognized, it will only depend on the data set provided in the training data - the training program (program 1) should detect how many patterns are in training file, and the recognition program (program 2) should use that dataset to recognize patterns in the out-of-sample data. Pattern names must not be hardcoded anywhere in the program code, including training program (program 1) or pattern recognition program (program 2) - training data will be provided, and out-of-sample data will be provided to test the programs - the data is a time series representing a variety of price points of a financial instrument, z-score method must be used to normalize the data before the data is used for training - name of patterns to be recognized will come from the training data file - the program that recognizes patterns based on trained data will provide the following output, pattern names used below are examples only: - if pattern A is found then say "pattern A found" - if pattern B is found then say "pattern B found" - if neither pattern A nor B are found, say "No pattern found" - this will be done for each set of data points from the out-of-sample file - you are free to use any Machine Learning model you find most suitable but must be supervised learning, and effective in recognizing patterns in the financial data time series (like CNN, CNN + LSTM) - use z-score normalization to normalize data (both in-sample data before training the model, and out-of-sample data before running pattern recognition routine). This can be done on the fly and kept in the memory, no need to save normalized data to a file - you must poses good understanding of Machine Learning domain knowledge, including NN training, normalization and regularization, and be able to explain the code, and advice on best approach - you must be proficient in Python and public domain Machine Learning libraries - Machine Learning Python libraries used must be public domain