Sberbank presented its new platform GigaCode, based on large language models, which helps IT specialists to increase the efficiency of the development process. Sberbank representatives informed CNews about it.
At the moment, several thousand specialists are already using GigaCode, and in the coming months the platform will become available to external developers as well. This was announced by Andrey Belevtsev, Senior Vice President, CTO and Head of Technology at Sberbank.
GigaCode has the ability to generate the most probable continuations of a code fragment right in the development environment in real time. The platform also offers other features such as advanced code generation, semantic code search, search for certain types of vulnerabilities, etc.
GigaCode is co-developed with GigaChat using similar technologies, models, processes and approaches. GigaCode is based on neural network models trained on large amounts of source code and specialized for specific development tasks.
The platform supports popular programming languages such as Java, JavaScript, Python, TypeScript, C/C++ and others, and is compatible with various development environments including IDEA, PyCharm, VSCode, Jupiter and others.
Andrei Belevtsev, Senior Vice President, CTO, Head of Sberbank's Technology Block, notes that the Russian market responds quickly to new challenges and offers effective solutions. He emphasizes that the GigaCode tool can become a reliable assistant for developers, helping to increase their productivity. There are many routine operations in programming, such as writing similar code fragments that are performed mechanically. Belevtsev is confident that GigaCode will simplify programmers' lives, allowing them to focus on key tasks and automate the algorithmic aspects of their profession.
Ailib neural network catalog. All information is taken from public sources.
Advertising and Placement: [email protected] or t.me/fozzepe