Robotics application development architecture

A system, comprising:

• one or more processors;
• and
• memory that stores computer-executable instructions that, if executed, cause the one or more processors to:
• obtain a first set of parameters of a robotic device associated with a customer account and a second set of parameters specifying a simulation environment for testing an application of the robotic device, the first set of parameters indicating a data storage location of the application and the customer account and the second set of parameters indicating a selection of the simulation environment from a set of simulation environments;
• obtain a reinforcement learning model for the application and a termination condition for completing training of the reinforcement learning model, wherein the reinforcement learning model comprises a customer-defined reinforcement function;
• select, from a pool of resources, a set of resources on which to execute a set of simulations in the simulation environment, the set of resources selected based, at least in part, on sufficiency of the set of resources for executing the set of simulations and training the reinforcement learning model;
• obtain the application from the data storage location;
• load the application onto the set of resources;
• execute the set of simulations with the application to train the reinforcement learning model until the termination condition is reached;
• monitor execution of the set of simulations to obtain data indicative of performance of the application;
• and
• provide the data to enable modification of the application and the reinforcement learning model during the execution of the set of simulations, wherein:
• the modification of the reinforcement learning model comprises receiving, from a customer, an updated customer-defined reinforcement function, and
• the modification of the application is based on at least one of the performance of the application, the updated customer-defined reinforcement function, and the termination condition.

A non-transitory computer-readable storage medium comprising executable instructions that, as a result of being executed by one or more processors of a computer system, cause the computer system to at least:

• obtain a first set of parameters of a robotic device associated with a customer account and a second set of parameters specifying a simulation environment for testing an application of the robotic device;
• obtain a set of reinforcement learning models, wherein a reinforcement learning model of the set of reinforcement learning models comprises a customer-defined reinforcement function;
• select, from a pool of computer system instances, a set of computer system instances on which to execute a set of simulations in the simulation environment, the set of computing system instances determined to be sufficient for executing the set of simulations and training the set of reinforcement learning models;
• execute the set of simulations in the simulation environment, each simulation simulating a reinforcement learning model of the set of reinforcement learning models;
• obtain a selection from the set of reinforcement learning models indicating a selected reinforcement learning model;
• update the application based on the selected reinforcement learning model;
• select a fleet of robotic devices from a set of robotic devices associated with the customer account;
• transmit, to the fleet of robotic devices, associated with the customer account, that includes the robotic device, a resource identifier indicating a location of the application data to cause the fleet of robotic devices to install and execute the application;
• obtain, from the robotic device, data generated in response to sensor inputs obtained by the robotic device;
• and
• update, using the data, a graphical user interface to enable monitoring of the robotic device and updating of the customer-defined reinforcement function with a customer-defined update during execution of the application.