Every few months, the CircleCI product organization is going to round up some of the cool stuff we’ve created for you. Read on to learn about the new security, productivity, and speed features you can take advantage of.

In Agile development, where development is repeated in short periods, the key unit of the development cycle is called an iteration. Iterations, consisting of Design, Development, Testing, and Improvement are usually set for 1 to 4 weeks, and they are characterized by completing a full cycle of system development. After completing one cycle and releasing it, known as Iteration 1, the process is repeated with Iteration 2, Iteration 3, and so on.the.

The power of LLMs to solve real-world problems is undeniable, but unfortunately, in some cases, only theoretical. What’s stopping us from getting the most out of OpenAI’s text completion capabilities in production apps? One common problem is the inability to confidently guard against bad outputs in production the way we’re used to doing with non-AI test suites. Let’s go one step deeper. There is no equivalent of code coverage for an LLM.

The testing pyramid model untangles the complexity of software testing by fitting it into an efficient hierarchical structure. By focusing on unit tests at the base, integration tests in the middle, and end-to-end tests at the top, the testing pyramid ensures that most testing efforts are spent on tests that are fast, reliable, and easy to maintain. This allows for quicker iterations, improved code quality, and more stable releases.

An LLM hallucination occurs when a large language model (LLM) generates a response that is either factually incorrect, nonsensical, or disconnected from the input prompt. Hallucinations are a byproduct of the probabilistic nature of language models, which generate responses based on patterns learned from vast datasets rather than factual understanding.

Microservices architecture is a method of developing software systems that structures an application as a collection of loosely coupled services, each focusing on a single function or business capability. Each service operates within a discrete, confined context, communicating with other services through well-defined interfaces — typically APIs.

Test-driven development (TDD) is a software development process that involves writing tests for your code before you write the code. This approach has transformed the development methodology around testing. While the traditional waterfall model of software development was linear, with testing occurring near the end of one long timeline, TDD makes testing an ongoing, iterative process.

Prompt engineering is the practice of crafting input queries or instructions to elicit more accurate and desirable outputs from large language models (LLMs). It is a crucial skill for working with artificial intelligence (AI) applications, helping developers achieve better results from language models. Prompt engineering involves strategically shaping input prompts, exploring the nuances of language, and experimenting with diverse prompts to fine-tune model output and address potential biases.

PyTorch is an open-source machine learning (ML) framework that accelerates the path from research prototyping to production deployment. You can work with PyTorch using regular Python without delving into the underlying native C++ code. It contains a full toolkit for building production-worthy ML applications, including layers for deep neural networks, activation functions and optimizers. It also has associated libraries for computer vision and natural language processing.

CircleCI jobs can use OpenID Connect (OIDC) identity tokens to securely access cloud providers without having to store a static credential in CircleCI. This article describes how to access Azure resources using OpenID Connect (OIDC).