Cloud native technologies have been observed to expand into the realm of Internet of Things (IoT) and Cyber-physical Systems, of which an important application domain is robotics. In this paper, we review the cloudification practice in the robotics domain from both literature and industrial perspectives. We propose RoboKube, an adaptive framework that is based on the Kubernetes (K8s) ecosystem to set up a common platform across the device-cloud continuum for the deployment of cloudified Robotic Operating System (ROS) powered applications, to facilitate the cloud native evolution in robotics. We examine the process of modernizing ROS applications using cloud-native technologies, focusing on both the platform and application perspectives. In addition, we address the challenges of networking setups for heterogeneous environments. This paper intends to serves as a guide for developers and researchers, offering insights into containerization strategies, ROS node distribution and clustering, and deployment options. To demonstrate the feasibility of our approach, we present a case study involving the cloudification of a teleoperation testbed.

Remarkable performance of large language models (LLMs) in a variety of tasks brings forth many opportunities as well as challenges of utilizing them in production settings. Towards practical adoption of LLMs, multi-agent systems hold great promise to augment, integrate, and orchestrate LLMs in the larger context of enterprise platforms that use existing proprietary data and models to tackle complex real-world tasks. Despite the tremendous success of these systems, current approaches rely on narrow, single-focus objectives for optimization and evaluation, often overlooking potential constraints in real-world scenarios, including restricted budgets, resources and time. Furthermore, interpreting, analyzing, and debugging these systems requires different components to be evaluated in relation to one another. This demand is currently not feasible with existing methodologies. In this postion paper, we introduce the concept of reasoning capacity as a unifying criterion to enable integration of constraints during optimization and establish connections among different components within the system, which also enable a more holistic and comprehensive approach to evaluation. We present a formal definition of reasoning capacity and illustrate its utility in identifying limitations within each component of the system. We then argue how these limitations can be addressed with a self-reflective process wherein human-feedback is used to alleviate shortcomings in reasoning and enhance overall consistency of the system.

When our small Emergency Radiology lab sought to engage in AI research, we found that we lacked needed resources, and pre-existing AI research systems did not translate to our workflow or adapt to our needs. Without a system to manage the many facets of setting up and performing AI research, significant manual efforts and a constellation of incongruent tools are needed. A wide range of effort-intensive operations combined to make AI research infeasible for us: Data curation, annotation, machine learning model development, management of people and resources, security, auditing, and multi-system interoperability are far too large of a simultaneous undertaking for a resource-limited lab to manage. The costs of a large staff and requisite resources to perform all of these activities were prohibitively high. In order to perform rapid research, development, and deployment of AI models with minimal staff and low-cost resources, we needed a system that could orchestrate all of these necessary tasks, without the omissions, gaps, and incongruities between tools that so often require many resources and manual intervention. We set out to design an integrated platform that could facilitate the plurality of our research initiatives. Our goal was to restore radiologists as the drivers of innovation in imaging-focused AI. Our design philosophy was that tasks that could be automated, such as handling, translating, and curating high-quality data, should be handled by computers rather than armies of annotators, data scientists, and engineers. The hurdles to successful facilitation of imaging machine learning have been well documented [1].

A data orchestrator acts to integrate and fine tune its instruments. It is agile, scalable and evolutive to adapt quickly. Data orchestration delivers the core from which people, processes and systems and data unite to provide one source of the truth. It is not a matter of if and when – but how – to regain valuable time while increasing profit. Up until now, the visionary leaders who have understood the importance of integrating reliable data to validate decisions have incrementally introduced various tools from vendors offering singular solutions for individual functions of the business. It is through the narrowed focus of these early disruptors we now know that addressing bit parts of an organization, as unique products come to market, is not delivering the gains they expected and often becomes a binding decision to commit to one provider.