Sai Kumar Arava on the Applications of Machine Learning

by William Jones / ⠀AI / October 19, 2024

From securing admission to the prestigious IIT Kanpur to leading AI innovations at Adobe,

Education and Early Impact

Sai Kumar’s academic journey is marked by excellence. He holds a dual Master’s degree in Mathematics and Computer Science from IIT Kanpur, providing the perfect foundation for a career in artificial intelligence. During his studies, he conducted preliminary research on machine learning applications, showcasing early promise. Right out of college, Arava landed a position at PayPal, where he developed a system for tracking behavioral insights to enhance user experience. Recognizing the potential of analytics, he transitioned to Adobe.

Breakthroughs at Adobe & Beyond

At Adobe, Arava spearheaded the “Intraday Modeling to Adjust Online Ad Distribution,” which significantly improved the accuracy of machine learning models by considering intraday campaign traffic. The Click models developed within this framework have had a significant financial impact, contributing millions in revenue for several customers. Relocating to the United States, Arava focused on algorithmic marketing attribution. He productionized discrete-time survival-based techniques for algorithmic marketing attribution that were patented and cited by many big tech companies for their own models. His work at Adobe leveraging marketing attribution and deep learning is patented and widely cited more than 100 times by top companies like Google and Meta. The fact that his technology has been embraced and incorporated by such a diverse range of companies, including e-commerce giants, technology powerhouses, and transportation innovators, underscores the broad applicability and relevance of the models developed within this framework. This cross-industry recognition and adoption demonstrate the far-reaching impact and importance of the work. This cross-citation suggests that the underlying technology and methodologies have been recognized and built upon by industry leaders, further highlighting the significance and impact of the work. Arava took on the management role to lead the research, engineering, and product for marketing attribution that helped launch product in a short time. Arava says launching a marketing attribution product at Adobe and having seen it used by several Fortune 100 customers for measuring their marketing performance for their new product launches affecting 280 million users internationally and resulting in millions of revenue is one of the most satisfying parts of his career. His models have empowered marketers to optimize their campaigns, resulting in a 20% increase in customer acquisition for a leading streaming company.

Leading AI Innovation Today

Sai now leads a large team of machine learning engineers for B2C and B2B AI intelligent services that help businesses target audiences more effectively by providing propensity scores for B2B and B2C businesses. Businesses can add personalization based on their goals thereby saving millions of ad dollars in unnecessary and inefficient targeting and saving users from email fatigue. With the increasing usage of Generative AI across all enterprises, Sai Kumar’s current research focuses on improving the capabilities of large language model-based marketing analytics copilots. His work involves innovative approaches combining semantic search, prompt engineering, and fine-tuning to enhance the accuracy and reliability of AI models in executing domain-specific tasks related to marketing mix modeling and attribution and predictive machine learning services.

The problems in B2B and B2C

“I believe AI has the potential to revolutionize how businesses understand and interact with their customers. My goal is to make that potential a reality”. Arava says. Many businesses struggle to identify campaigns that will resonate with their audiences, especially when it comes to personalization and recommendations. With countless potential customers, it can be difficult to target the right keywords, outlets, and strategies. Instead of doing this randomly, companies may find more success with machine learning models. “They must learn from patterns to identify people with similar behaviors,” Arava says, adding that the “user experience will be much more mature” as a result. Arava says that the machine learning field is fast-paced, and as a result, there is a lot to learn, as well as a lot of demand from the industry. This is because the skills required to make an impact include a background knowledge of math and behavior analytics. Arava says that “keeping up is a challenge, but also worthwhile.” And with the explosion of big data, Arava says that marrying the two subjects is new for not just the industry, but himself.

Improving businesses with advanced analytics

Sai Kumar Arava is at the forefront of revolutionizing marketing analytics with AI, driving millions in revenue and transforming how businesses understand their customers. During his time, he has seen machine learning and data improve businesses. He has also learned how to pass the end value to customers and offer unique insights. “My expertise has helped launch several key projects that are delivering millions of dollars in revenue for large enterprise businesses,” Arava says. “In the last few years, I was able to attend, network, and speak at major machine learning conferences like Knowledge Discovery and Data Mining, which helps me share my research ideas with the community,” Arava says. “My work in deep learning in multi-touch attribution has been recognized internationally by several other researchers building upon my work.”

Aspirations for the Future and Leadership

Arava says that he is currently working on a challenging AI project that requires a combination of the latest generative AI Technology. This project could change how enterprises can deliver or use data or retrieve insights from their data. “I’m excited about this project,” Arava shares. “We could see multiple orders of magnitude in terms of value realization and efficiencies, potentially changing how businesses operate in the next few years.”