Careers & Student Outcomes

From Stevens to Peloton: Aaryan Yekkisetty’s Journey into the Future of Fitness Tech

Recent Machine Learning graduate student turning natural language processing passion into action with fitness giant in New York City

When Aaryan Yekkisetty arrived at Stevens Institute of Technology to pursue his master’s in Machine Learning, his mind was already two steps ahead, peering into what the future could entail at the end of his studies. Before setting foot on campus, he had reached out to professors, eager to dive into research and explore the frontiers of artificial intelligence. That proactive mindset would come to define his time at Stevens — and ultimately help him land a coveted role as a machine learning engineer at Peloton in New York City.

"Stevens stood out as the perfect place to pursue my master’s in Machine Learning," said Yekkisetty, who received his Master of Science this spring. "The curriculum was well-balanced with four core foundational courses and a wide array of electives that let me explore both the breadth and depth of AI. Stevens had the academic rigor, supportive faculty, and prime location right next to New York, which meant I could immerse myself not only in studies but also in the thriving tech and startup ecosystem just across the river."

In the Department of Computer Science, Yekkisetty found more than just coursework. He found mentors who shaped his thinking and projects that pushed the boundaries of what AI could do. He credits Assistant Professor Nikhil Muralidhar’s "Mathematics for Machine Learning" course for deepening his theoretical understanding, and Teaching Associate Professor in the Department of Systems and Enterprises Carlo Lipizzi’s guidance on his master’s project — focused on energy optimization for large language models (LLMs) — for teaching him the art of impactful research.

He also worked closely with Assistant Professor Zining Zhu on autonomous agents using vision-language models for underwater navigation, and served as a teaching assistant for both Muralidhar and Joseph Helsing.

"Teaching gave me a hands-on chance to help newcomers build strong coding habits, and it also taught me the importance of being able to teach and communicate complex logic clearly," said Yekkisetty.

Stevens stood out as the perfect place to pursue my master’s in Machine Learning. The curriculum was well-balanced with four core foundational courses and a wide array of electives that let me explore both the breadth and depth of AI. Stevens had the academic rigor, supportive faculty, and prime location right next to New York, which meant I could immerse myself not only in studies but also in the thriving tech and startup ecosystem just across the river.
Aaryan YekkisettyM.S. '25 in Machine Learning

Networking and Natural Language Processing Lead Way to Peloton

Aaryan Yekkisetty poses with a metal plaque representing the Peloton logo outside Peloton's New York City office.Aaryan Yekkisetty works for Peloton’s NLP/NLU team in New York City.

Outside the classroom, Yekkisetty was just as driven. He attended AI meetups, hackathons and tech talks across New York City, building a network and staying on the cutting edge of machine learning.

"I spent every free moment focusing on three things — applications and networking, reading the latest in NLP (natural language processing), and doing logic-building with LeetCode," said Yekkisetty. "Stevens gave me strong AI fundamentals, so I was able to focus more on deepening my specialization and sharpening my practical skills."

Yekkisetty also credits Stevens for helping him land an internship with Con Edison, where he created tools that automate password security and AI-based troubleshooting for smart meters. The experience gave him his first taste of how impactful AI can be in enterprise systems.

That hustle paid off. After interviewing with more than 20 top AI teams with industry titans such as Apple, Amazon and Capital One, Yekkisetty found his perfect match at Peloton, a company that creates at-home exercise bikes that leverage technology to create a communal feel without going to the gym. "I’ve always been into fitness, and Peloton’s mission to make fitness more immersive and personal through AI deeply resonated with me."

Now part of Peloton’s NLP/NLU (natural language understanding) team, Yekkisetty is working on upgrading the company’s speech-to-action pipeline, enabling users to issue real-time voice commands during workouts. He’s also exploring multilingual translation systems to make Peloton’s content more accessible to a global audience.

"The best part is seeing my work in action," he said. "Every model I contribute to directly improves someone’s workout experience. It’s incredibly fulfilling to know that what I build helps people stay motivated, healthy and connected to their fitness journey."

The recruitment process at Peloton was rigorous — six rounds covering everything from deep learning and NLP to system design and coding. But Yekkisetty stood out thanks to his alignment with the role and his unique blend of experiences. His work at Con Edison, where he built AI tools for smart meter troubleshooting, and his contributions to OpenAI’s Whisper speech-to-text model gave him a strong edge.

Yekkisetty also brought a passion for open-source and innovation. One of his proudest moments was winning first place in the Google Track at Columbia’s DivHacks Hackathon for "Vocal Vision," a project that turned educational content into lip-synced video using AI.

Looking ahead, Yekkisetty is excited to deepen his expertise in NLP and multimodal systems and eventually explore the intersection of AI and robotics. Whether he stays in industry or launches a startup of his own, one thing is clear: Yekkisetty is just getting started.

"I see AI as a way to reimagine what’s possible. I plan to keep building toward that future, one model and one idea at a time," said Yekkisetty.

Learn more about academic programs and research in the Department of Computer Science: