Leading artificial intelligence (AI) platform companies are developing AI portfolios and APIs to allow AI engineers to build specialized enterprise AI applications with foundation models. The companies’ products help AI and machine learning (ML) engineers employ techniques to take large data sets from disparate sources and merge them together to gain valuable business insights. AI engineers can also customize the provided AI models with proprietary data to drive tasks from product design to drug discovery. Here, we look at some of the top companies that are supporting AI engineers in developing enterprise AI applications.
Google offers a variety of AI products and tools to aid software engineers in creating software: such as Google AI for Developers, which provides generative AI models and tools; Google Could Vertex AI, a platform for building and scaling applications; Android for Developers, for mobile phone applications. Google AI Studio allows users to take advantage of the company’s Gemini ecosystem, which supports AI products, platforms and APIs. It offers products to generate emails, create and edit photos and videos and provide automatic dubbing. The company says its AI offerings can help companies boost their productivity and streamline their operations.
IBM’s AI line for software developers includes watsonx, a platform that allows users to train, tune and distribute models with generative AI and machine learning capabilities. The platform allows users to train, validate and deploy ML models, scale workloads to handle their data and create data and AI workflows that are transparent and explainable. It enables a company’s employees to create virtual assistants without having to write code and also provides developers with recommendations to help them write code.
Microsoft offers Azure AI Studio, a development platform companies can use to select and deploy the AI models for their needs. The studio provides management for data preparation, model development and training and provides enterprise support for ML libraries Pytorch and TensorFlow. It provides APIs, such as OpenAI Service, and has introduced Copilot and GitHub Copilot as AI engineer tools.
NVIDIA offers AI tools that it says can speed up the AI workflow while increasing efficiency and lowering costs. The company provides solutions for generative AI, data analytics and inference models. NVIDIA’s Metropolis application framework applies ML to visual data from internet of things (IoT) devices to handle retail, inventory management, traffic engineering, optical inspection in factories and patient care. The framework offers real-time threat detection for improved cybersecurity. It allows users to build conversational applications, including translation.
Meta is performing research in a number of areas of AI development. For instance, it is exploring how to use computer vision to provide machines with a better understanding of the world. The company is working to improve ML by building algorithms that are compatible with and inspired by human cognition. Meta is developing agents that can be embodied in robots to interact with people in both physical and virtual environments. It is also working to expand the capabilities of generative AI. Among its development products are Llama, its open-source large language model (LLM), and Cicero, an AI that can play the game Diplomacy at a human level.
Amazon is making AI accessible to data scientists, business analysts and students, among others, through its AWS unit. AWS provides foundation models to help AI software engineers build their own applications powered by generative AI. The company provides tools and services, such as Amazon SageMaker, which lets users build, train and deploy ML models. It also supplies frameworks, such as Hugging Face, Pytorch, TensorFlow and Jupyter, optimized for AWS as well as provides infrastructure and training to support ML projects.
Oracle offers AI engineering capabilities through its data service and cloud platform. The company provides applications with embedded AI to support business functions. Users can choose between managed open-source or proprietary large language models and fine-tune them, augmenting them with their own business data. Prebuilt AI models can be customized with a company’s data and incorporated into its applications. It also offers cloud-based infrastructure, such as GPUs, to support customer ML and AI activities.
Databricks offers AI developers its Data Intelligence Platform that allows companies to use their own data with AI. The platform is built on a “lakehouse” architecture that combines attributes of data warehouses, large repositories of structured data, with data lakes, which store raw data in a variety of formats. The company says its platform can automatically optimize performance and manage infrastructure in a way best suited to a particular customer. The platform also integrates with a company’s existing tools for AI, business intelligence (BI) and governance.
Anthropic makes a family of Claude AI models, which it says go beyond pattern recognition to perform complex cognitive tasks. The model can analyze images, including handwritten notes, graphs or photos. It provides users with code generation capabilities, allowing them to create websites with HTML and CSS, turn images into structured JSON data and debug complex code bases. The company has teams researching how to make large language models more interpretable, how to keep AI helpful and how AI can be designed to have positive impacts on society.
OpenAI’s GPT model enables AI engineers to accept any combination of text, images, video or audio as input and produce any combination of text, image and audio output. The model can be customized to a company’s workflow and allow for collaboration among teams. It is being used to develop applications across corporate functions, including engineering, sales, finance, IT and human resources. The company believes its research will eventually lead to artificial general intelligence (AGI) or AI systems that are "generally smarter than humans."
