![\"\"](\"https:\/\/miro.medium.com\/v2\/resize:fit:640\/format:webp\/1*hiQY0y6x-tijAtF1Zqcpdg.jpeg)
<\/picture><\/div>\n<\/div>\n<\/figure>\nEarlier this year, leaders in our Adobe Romania office in Bucharest engaged our teams in a program designed to provide guided opportunities for out-of-the-box thinking, design, and innovation separate from their \u201cday jobs.\u201d Called \u201cGarage Week,\u201d this experiential program allows for broad experimentation from our security teams and other local teams to try and solve particularly persnickety problems they have always wanted to work on. We took an opportunity to sit down with Alexandru Calistru<\/a>, manager of our EMEA Security Coordination Center (SCC) and one of the curators of this year\u2019s Garage Week for a Q&A to learn more about the event and its goals.<\/p>\n Garage Week is a week-long event filled with workshops designed to encourage free thinking, innovation, and knowledge sharing on all of the great projects that come out of the week. The projects can be related to one\u2019s day-to-day job or they can be personal passion projects. The program offers great opportunities for collaboration and brainstorming among our teams with a pure focus on innovation. All employees in our Bucharest office, regardless of their role, are encouraged to participate in the program.<\/p>\n One key example of a benefit to the rest of Adobe came out of our most recent Garage Week: a new AI\/ML project the team called \u201cSherlOCk.\u201d A project initiated by one of our interns, Radu Chivereanu<\/a>, under guidance from Tiberiu Boros<\/a>. Project SherlOCk helps automate the information discovery process using Slack-based \u201cquestion bots\u201d that help users get more natural and quicker access to frequently asked questions. This project is an example of how our security team is investing in AI-based tools to help improve our processes and save our team member\u2019s valuable time.<\/p>\n Another project that developed out of this effort uses residual neural networking techniques. This project enables deep learning models with tens or hundreds of layers that train easily and approach better accuracy when going deeper into security intelligence data sets. This will help us design controls and tooling to better thwart potential adversary behaviors. In addition, the teams experimented with multi-modal AI models \u2014 models that combine static images, video, and audio \u2014 to better understand their behavior and potential use.<\/p>\n One of our goals this year was to work with our incident response teams to help develop better tools to help improve our triage time. Project SherlOCk\u2019s AI model can help our incident response teams more quickly understand whether a particular activity is malicious or not. Our incident response teams can query our available threat intelligence sources and receive more insightful responses that help teams more quickly understand whether a particular activity is malicious or not. This capability can help reduce the overall time of investigations by several orders of magnitude.<\/p>\n We continue to improve collaboration between our security teams globally. Activities like Garage Week help give time, resources, and creative space for teams to work together on innovative security projects that can help improve Adobe\u2019s overall security posture and reduce the potential attack surface across our products and infrastructure. We are working with other security leaders across our organization to develop and expand programs like this to help further foster creativity and innovation.<\/p>\n We would like to thank Alexandru for taking the time to chat with us about the Garage Week program from our Romanian team. Look out for more stories about our programs that encourage Adobe\u2019s culture of creativity throughout our security teams here on our blog.<\/p>\n Your opinion matters to us. Help shape the future of our blog by sharing your ideas and preferences. Click the link below to take a quick survey and tell us what you\u2019d like to read about next.<\/p>\n > Take the Security@Adobe Tech Blog Survey<\/strong><\/a><\/p>\n<\/div>\n