Integrating MultiTask Framework to Predict Road Data

My team and I are working on creating a Multi-Task ML Model that can predict road meta data (max speed, number of lanes, one-way or two-way, etc.) I would like to know how I can integrate the model or the predicted data into OSM to be able to deploy it as a public service. We are focusing on the Qatar data since that is where this project is taking place, but we are also testing it on parts of Asia and Europe to make sure it is applicable worldwide. Please let me know how to proceed with integrating our model - I prefer a fast solution since our project is coming to an end within a month.

Not possible to “integrate” “fast” directly in 1 month. That’s not how it should work. Even assuming expertise in OSM (which I frankly assume you all don’t have from this question), you need to sort out your license, prove your data quality, document your methodology, discuss with the local and global communities for at least 2 weeks to settle any disagreements, and do conflation to combine only more accurate data from you with existing ones. Import/Guidelines - OpenStreetMap Wiki
It’s not something inexperienced users should do, for good reasons. Very high risk to go wrong easily causing a mess to clean up afterwards, with potentially less reward than you think.

The project is taking place within a research institute as part of an internship, and we plan to publish a research paper with all the mentioned points (methodology, etc.) We have been researching OSM for almost a month now and are experimenting with different data types until we find the optimum one. This inquiry was posed in hopes of benefiting others with our research once it is done.

Other than releasing the dataset, could we not release a model that helps predict the datasets? It is not meant to replace already existing data, but rather has a public service to be able to correctly predict the road data. @Kovoschiz