Learning to Tackle Ambiguity

Shikher Srivastava (MSAI '23) was in his third year of studies at the Vellore Institute of Technology in India when he became fascinated with data science. He was tasked with analyzing a dataset of his choosing, and he focused on Twitter (now X) sentiment about demonetization in India.  

He collected thousands of tweets, cleaned the text data, built classification models, and extracted insights about public opinion.   

"Seeing how you could take unstructured social media data and turn it into meaningful analysis got me hooked," Shikher said. "That project was really the turning point for me."

That fascination led him to learn more about data models, which introduced him to AI and machine learning. He spent three years at financial services organization HSBC, where as a software engineer he worked on mobile banking platforms and building conversational AI features. He realized he needed to understand AI better in order to stay current with cutting-edge research. 

For that, he turned to Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program. 

"MSAI struck the right balance with strong technical coursework, deep learning, generative models, computer vision, natural language processing, combined with actual industry projects," Shikher said. "The program taught me not just how to build models, but how to integrate them into workflows, collaborate with stakeholders, and communicate findings effectively." 

Today, Shikher applies those lessons to his work as a senior data scientist at Affinity Solutions, a consumer purchase insights company. In his role he builds systems that help the company make sense of consumer transaction data. His work spans the full analysis process, from building machine learning models to getting them to run in production.  

The end goal is to take raw credit and debit card data and turn it into insights that help the company's customers understand spending patterns and market trends.  

"We're classifying billions of transaction records with new data coming in daily, making sense of it to identify merchants, locations, purchase patterns, and more," he said. "It's technically challenging work that requires pushing the boundaries of automation and applied AI at scale." 

Beyond the scale of the work, it's the impact that motivates Shikher. 

"These systems create tangible value. Our models unlock marketing insights that help brands understand how consumers actually spend and make smarter business decisions," he said. "That combination of hard technical problems and real-world impact is what keeps me excited." 

It was in MSAI where he realized how AI can deliver impact, and not just metrics. He also credits the program with teaching him the importance of systematic experimentation, as well as how to think about broader concepts related to AI, such as data quality, ethical implications, and long-term maintenance.  

Shikher completed his MSAI practicum project with information technology solutions company CDW and his capstone with Vanguard, one of the world’s leading investment management firms. Both experiences taught him valuable communication skills he relies on daily, whether it's working with product teams or presenting to leadership.  

"MSAI gave me the technical foundation I use every day," he said. "I also learned to translate technical work into business value and understand what questions matter to different audiences." 

More than anything else, MSAI helped Shikher develop a mindset that he believes is instrumental to his growth and professional success.  

"MSAI gave me confidence to tackle ambiguity. The field of AI moves incredibly fast, and MSAI taught me how to learn continuously and stay current with new techniques and approaches," he said. "Real-world problems rarely come with clean datasets or clear objective functions, and the program taught me to break down complex problems systematically and find practical solutions."