Post-Training and Deploying Open Source Reasoning Models in Foundry | BRK232 2mo Microsoft foundry Open Source reasoning models Reinforcement learning
