“In June, I left Amazon to join Meta as a machine learning software engineer for a total compensation of over $400,000. I was really excited about it and knew I wanted to take the job as soon as I got the offer,” Tumu said, ET reported, quoting Business Insider. Explaining his decision, he added, “Though I had learned a lot at Amazon, I just thought there was more interesting work going on at Meta.”

Real-world experience over personal projects

Tumu highlighted the importance of internships and practical experience over personal projects on resumes. He advised students to pursue internships during college, even if they are low-paying.

“When I see the resumes that people post online, asking for advice, I see projects or programming languages taking up space,” he said. “My general advice would be that once you have two or three years of experience, it's OK to remove the projects and focus more on highlighting your experience,” ET quoted Business Insider.

He added that he did not rely on referrals but applied directly through company websites and LinkedIn, emphasizing the importance of a strong resume to stand out.

Interview preparation is key

Tumu stressed that behavioural interview preparation is critical, calling lack of preparation one of the biggest mistakes candidates make. His interviews at Amazon and Meta included coding, machine learning, and behavioural rounds.

“I studied the company's values to prepare for the behavioural interview and had a huge document where I would write down stories to answer possible questions along with follow-ups,” he said, ET quoted Business Insider.

Early career choices shape the path

Tumu admitted missing out on internships during college but later secured a contract role after graduation. When choosing between traditional software engineering and machine learning, he opted for the latter despite lower initial pay because of his passion for the field—a decision that eventually led to his role in Meta’s advertising research team.

AI and machine learning evolution

Reflecting on the field, Tumu noted the shift from classical techniques to deep learning in artificial intelligence. “It used to be a lot more acceptable to just use classical techniques, which rely on humans to make decisions about data representations. Now the focus is on deep learning, which taps into artificial neural networks to automatically learn features from raw data,” ET quoted Business Insider.

He also highlighted that job titles in AI vary widely across companies, ranging from research scientist to applied scientist, software engineer, or machine learning engineer.