Embracing my inner ‘mad scientist’ helped me get an AI job at Adobe<\/p>\n
I went to art school in the US and didn’t have a formal machine learning background. I took my AI education<\/a> into my own hands, learning as I went through YouTube videos and friends.<\/p>\n Instead of trying to learn everything about AI at once, I focused on experimenting with Cursor<\/a> with my friends, who could hold me accountable. I treated this process like a formula: one hour of Cursor a day + seven days = a functional AI prototype.<\/p>\n Kumar learned AI skills from YouTube and friends.\u00a0<\/p>\n Courtesy of Nitya Kumar<\/p>\n I love dance, and by the end of the seven days, I had created a fun game that could detect and track the gestures of a user dancing in their room. It helped me develop a vibe coding workflow<\/a>, which involved using different AI tools for different processes. Though I started out with Gemini and Cursor, I gradually gained more confidence with other tools, too, like Figma MCP.<\/p>\n Instead of just building safe case studies with AI, I unleashed my inner “mad scientist.” For example, I created a tool that generated Matcha recipes. I used Gemini to develop and refine prompts I’d feed into Claude to vibe-code the product.<\/p>\n These quirky experiments helped me gain a grasp of the AI landscape and the tools that exist, and, I think, helped me stand out in interviews. They made interviewers laugh, and we’d start debating about matcha flavors and play around with the tools I’d built.<\/p>\n<\/p>\n