Want to become an AI developer? Here’s how with MySQL HeatWave

March 13, 2024 | 5 minute read
Nipun Agarwal
Senior Vice President, MySQL HeatWave
Text Size 100%:

Generative AI is arguably the most important technology trend right now. It’s a transformative force that will have a significant impact on nearly every industry. This fast-moving disruption represents tremendous opportunities for developers to play a critical role in the evolution of almost every new and existing application. 

Indeed, OnHires notes that “the demand for developers for AI in Europe and the USA in the coming years will not only prevail over the demand for any other profession but will even significantly outpace hiring developers in other areas.”  They also underline that “The AI developer's salary is thought to be one of the highest in the IT area.

Becoming an AI developer means acquiring new skills, though that process can be straightforward if you’ve been developing MySQL-powered apps. MySQL has long been the most popular database for developers , and MySQL HeatWave makes it easy to build upon your existing skills to become an AI developer with in-database machine learning and support for generative AI and vector store. For instance:

•    MySQL HeatWave includes everything developers need to build, train, and explain machine learning models using data across both database and object storage. HeatWave AutoML automates the ML lifecycle including algorithm selection, intelligent data sampling for model training, feature selection, and hyperparameter optimization. No need to move data to a separate ML cloud service and no need to be an ML expert. You can build machine learning models using familiar SQL commands and HeatWave AutoML is also integrated with popular notebooks such as Jupyter and Apache Zeppelin.

•    We previously announced support for Generative AI with the MySQL HeatWave vector store. Currently in private preview, the vector store will enable you to leverage the power of large language models (LLMs) with your proprietary data to get more accurate and contextually relevant answers to your business than using models trained only on public data. With generative AI and vector store capabilities, you can build applications that interact with MySQL HeatWave in natural language and efficiently search documents in various file formats in the Object Store.

•    MySQL HeatWave is available on OCI, AWS, and Azure, providing you with great flexibility to build and deploy apps.

Vector Store and Generative AI enhancements for MySQL HeatWave 

Let us share more details about our vector store and generative AI implementation as well as the latest enhancements to help you better understand how easy it would be for you to become an AI developer using MySQL HeatWave.

•    MySQL Database and HeatWave now support a new, native VECTOR data type, enabling you to use standard SQL to create, process, and manage VECTOR data.

•    The vector store is integrated with MySQL Shell, allowing you to ask questions and receive answers in natural language using this familiar interface. The integration also allows you to query data using natural language via MySQL Shell for VS Code. Third-party tools such as Zeppelin and Jupyter can also access and leverage the GenAI and vector store capability of MySQL HeatWave.

•    The pipeline to discover and ingest proprietary documents in the vector store is automated, including transforming users’ unstructured data and generating embeddings, making it very easy for developers without ML expertise to leverage the vector store.

•    Vector processing is accelerated using the in-memory and scale-out architecture of HeatWave. The generation of embeddings in the vector store is parallelized across all cluster nodes, which means that multiple input files can be processed in parallel on multiple threads. As a result, ingesting unstructured data in various formats such as PDF, DOCX, HTML, TXT, and PPTX into the vector store is very fast

•    As a developer, you will have several options to use LLMs with MySQL HeatWave. You’ll be able to use the OCI Generative AI Service to access pre-trained, foundational models from Cohere and Meta, for example, to summarize, embed, and generate text. You’ll also be able to run in-HeatWave LLMs.

•    Another important benefit is that you can easily use generative AI jointly with other HeatWave capabilities, such as machine learning. For example, you could rapidly build an online food delivery application enabling customers to ask questions in natural language and get recommended dishes from various restaurants based on their preferences using the HeatWave AutoML recommender system.

HeatWave AutoML - Restaurant Use Case

Multiple use cases for AI involve analyzing data in various formats, stored across databases and object storage (fraud detection, clinical research, predictive maintenance, etc.). MySQL HeatWave Lakehouse allows you to query data in object storage, MySQL databases, or a combination of both with record speed. The query processing is done entirely in the HeatWave engine, enabling you to take advantage of HeatWave for non-MySQL workloads in addition to MySQL-compatible workloads. With that said, let’s review the recent enhancements to MySQL HeatWave Lakehouse.

MySQL HeatWave Lakehouse Enhancements 

We previously announced the following capabilities for the MySQL Database:
•    The ability to use HeatWave for real-time analytics on JSON documents stored in the MySQL database, accelerating queries on the documents by orders of magnitude.
•    Native support for JavaScript enabling developers to write stored procedures and functions in JavaScript and execute them natively inside MySQL HeatWave. Rich language features offered by JavaScript such as user-defined types, containers, and functional programming constructs simplify implementation vs. using SQL. Developers can write rich application logic in JavaScript and get high performance by executing the program inside the MySQL database.

These capabilities are now also available for data in object storage using MySQL HeatWave Lakehouse:
•    You can now use HeatWave to query semi-structured data in JSON format in object storage—for example, to develop content management apps or real-time dashboards using JSON data in object storage.
•    With native JavaScript support in HeatWave Lakehouse, developers can use JavaScript to process and query data in object storage, including for non-MySQL workloads. For example, you can use HeatWave Lakehouse to build dynamic content-loading applications using the rich features of JavaScript to process and query the data in object storage.

Run JavaScript with MySQL HeatWave in Object Store

AI is taking the world by storm and will have a significant impact across industries. With MySQL HeatWave, you can easily build upon your existing MySQL skills to become an AI developer and be part of this revolution. Get started today! Try MySQL HeatWave for free.

Additional resources

•    Learn more about MySQL HeatWave
•    Developer resources
•    Request a free MySQL HeatWave workshop

  i https://www.onhires.com/blog-post/the-most-demanded-ai-developers-in-the-world-and-their-salaries 
  ii https://www.jetbrains.com/lp/devecosystem-2023/databases/ 
  * Disclaimer: Benchmark queries are derived from the TPC-DS benchmark, but results aren’t comparable to published TPC-DS benchmark results since these don’t comply with the TPC-DS specification.

Nipun Agarwal

Senior Vice President, MySQL HeatWave

Nipun Agarwal is Senior Vice President of MySQL Database and HeatWave Development at Oracle. His interests include data processing, machine learning and cloud computing. Prior to this role, Nipun played a pivotal role in Oracle Labs, where he directed various research endeavors that later evolved into new Oracle products, such as MySQL HeatWave. Nipun's journey with Oracle began in 1994, following the completion of his Master of Science in Computer Science. For several years, he contributed to the Oracle database team. To date, Nipun has amassed an impressive portfolio of 200 patent awards.

Previous Post

Customer sentiment analysis with OCI AI Language

Aviv Graupen | 5 min read

Next Post

Introducing Zero to low-cost Autonomous Database for Developers

Simon Law | 3 min read