Upcoming features
In our continuous efforts to improve the Mercor platform, we are implementing many new and valuable upcoming features.
The backtesting functionality will allow you to test the performance of your algorithms on historical data.
The aim of the Mercor platform is to support AI-based algorithms. Thus, in the near future, the Mercor environment will support various machine learning and AI packages such as:
· Scipy.
· Scikit-learn.
· Theano.
· TensorFlow.
· Keras.
· PyTorch.
First of all, we want to make sure ‘regular’ algorithms function smoothly before implementing AI. To reach this goal your feedback is essential.
In order to create an environment in which the construction of algorithms happens as easyly and intuitively as possible, our team is planning on adding several tools to assist the construction and testing of an algorithm. In the current version, you will find a mock-up panel of tools that will soon be deployed. Some of the tools that will be included in coming updates:
· A python-based terminal to run and test functions
· A debugger
· Timing mechanism to optimize the speed of the algorithm
· Tutorials and documentation
· An overview of all third-party packages
· A forum
· The tools panelA git-based collaboration tool enabling teamwork
Your input is more than welcome, as we are building towards the ideal development environment. We would love to hear which tools you would like to see included, and we will gladly do our best to add them as soon as possible.
For all algorithms and especially AI-based algorithms, enormous amounts of data is essential for backtesting and training purposes. Mercor started collecting data from both public as well as private sources to offer exclusively to its developers.
To encourage the development of AI-based algorithms, we offer developers access to increased computational power in order to train their algorithms through our servers.
Last modified 1yr ago