{"id":11898,"date":"2024-08-09T12:43:42","date_gmt":"2024-08-09T19:43:42","guid":{"rendered":"https:\/\/intuitblog.com\/?p=11898"},"modified":"2025-04-22T06:31:10","modified_gmt":"2025-04-22T13:31:10","slug":"running-with-data-scientists-what-its-like-to-work-in-a-dynamic-tech-culture","status":"publish","type":"post","link":"https:\/\/www.intuit.com\/blog\/innovative-thinking\/running-with-data-scientists-what-its-like-to-work-in-a-dynamic-tech-culture\/","title":{"rendered":"Running with data scientists: what it\u2019s like to work in a dynamic tech culture"},"content":{"rendered":"\n

The work of data scientists isn\u2019t easily understood. Their work begins with laying a strong data foundation to perform robust data explorations followed by employing a set of mathematical techniques and experiments to deliver durable benefits. But it is not just the promising nature of their work. The environment in which they thrive empowers them to create meaningful experiences for customers and see impact in the business world. <\/p>\n\n\n\n

Driven by synergy<\/strong><\/h2>\n\n\n\n

The team of data scientists at Intuit have played a pivotal role in building AI-driven experiences, insights and recommendations to  customers long before generative AI (GenAI) and large language models (LLMs) became widely adopted by enterprise companies across myriad industries. The convergence of natural language processing (NLP), natural language understanding (NLU), machine learning, knowledge engineering\u2014and the new breed of GenAI\u2014are enabling the creation of virtual financial advice on a large scale, augmented by virtual financial experts in real-time. For many years, the team in India continues to build capability to generate meaningful financial insights. The key distinction between traditional AI technologies and GenAI-powered applications like Intuit Assist<\/a>  lies in how insights are triggered and presented to end users. <\/p>\n\n\n\n

Today, our financial technology platform produces 65 billion machine learning predictions daily, employs natural language processing to handle over 25 million conversations every year, and runs 2 million models in production per day for small business transaction categorization.<\/p>\n\n\n\n

\u201cWe create personalized customer experiences using commercial, open source and  our own custom-trained LLMs to orchestrate and decipher customer queries and preferences. What’s particularly exciting for us is the ability to utilize other cutting-edge LLMs in tandem with existing AI and data capabilities to provide superlative customer experiences,\u201d says Vignesh Subramaniam<\/a>, Intuit India AI and machine learning leader.<\/p>\n\n\n\n

These developments serve to supercharge the efficiency with which data scientists build models. \u201cOur team is creating a GenAI-driven system which provides personalized answers to financial questions for small business customers in QuickBooks, for example. They range from simple ones like \u2018what were my sales in the last quarter?\u2019 to complex ones like \u2018forecast my profit for the upcoming year and attribute reasons for it.\u2019 We combine the strengths of LLMs with classical algorithms to give answers that follow fluid natural language,” says Anusha Mujumdar<\/a>, senior manager in data science and AI at Intuit India.<\/p>\n\n\n\n

The team is able to generate algorithmic recipes, and automate and scale up insights in the blink of an eye. The interplay between different AI technologies allows for more sophisticated and nuanced analyses of the data, and highest standards of mathematical  and intellectual rigor within the team. The most rewarding part is the team builds on AI algorithms to generate recipes for computing statistics and automatically analyze data to generate insights and recommendations.<\/p>\n\n\n\n

Put algorithms to the test with actual customers <\/strong><\/h2>\n\n\n\n

Data science is a complex field rooted in mathematics and computer science, with no shortage of variables that can easily confuse even the most experienced practitioners. Certain patterns or correlations in the data may seem significant but could simply be coincidental or spurious.<\/p>\n\n\n\n

So, paying careful attention to data quality and integrity are critical components of the team’s approach, as they put their algorithms to the test. What this means is they collaborate more closely with designers and software developers to holistically understand customer experiences, inspired by a process known as a Follow Me Home<\/em><\/a>. <\/p>\n\n\n\n

\u201cWe provide customers with valuable financial insights that help our customers more deeply understand how to grow and scale their business,\u201d Anusha says. <\/p>\n\n\n\n

In practice, Follow Me Home allow team members to have conversations with customers in their home or office to learn about their experiences from using Intuit\u2019s AI-driven products. These observations help the team  to identify pain points, areas for improvement, and patterns that may not emerge through traditional data analysis. Such insights can be used to validate strategies, add depth and meaning to their data analyses.<\/p>\n\n\n\n

\u201cWe strive to create algorithms that work like magic. Customer obsession drives our team to push the boundaries of what we can achieve, far  beyond predetermined directives,\u201d Vignesh says. <\/p>\n\n\n\n

Break the mold<\/strong><\/h2>\n\n\n\n

The team not only values continuous learning, but also creates an environment where learning is  integral to the culture. This is evident in monthly paper clubs, where team members convene to read and discuss the latest research journals and algorithms. This practice encourages the team to stay up to date on trends,  techniques, and advancements in the field. <\/p>\n\n\n\n

On top of everything, they hold tech share-outs for each project, which are conducted in a collegial style, resembling PhD defense presentations. This type of learning raises the bar for data scientists and engineers to grasp sophisticated tech developments and gain valuable insights into how to improve ML and GenAI model development.<\/p>\n\n\n\n

\u201cPresenting papers at conferences is another significant milestone in our learning process. We won the Best Applied Track award for data scientists at CODS-COMADs in January 2024 for our paper on clickstream modeling framework for real-time customer event prediction,” Vignesh says. <\/p>\n\n\n\n

Meet the team of Intuit India data scientists<\/strong><\/h3>\n\n\n\n

Our team of exceptional professionals have distinguished themselves in three key areas: computer science, statistics, and optimization. Their expertise in computer science enables them to tackle complex problems, build efficient algorithms, and develop new solutions that leverage emerging technologies. The statisticians on the team possess the analytical skills necessary to make sense of complex data sets, derive predictive insights, and develop predictive models that enhance product performance. Members with optimization backgrounds solve problems by maximizing efficiency, reducing complexity, and identifying patterns that can enhance product functionality. <\/p>\n\n\n\n

Recently, Intuit India ranked first<\/a> among 50 best firms for data scientists to work for by Analytics India Magazine. This prestigious honor is a testament to our thriving culture and cutting-edge work that has been completed over time by diverse high-performance teams who pushed the boundaries of innovation in technology jobs<\/a>.<\/p>\n\n\n\n

Their commitment to solving customer problems is evident in this sampling of firsthand quotes from Intuit India data science team members.<\/p>\n\n\n\n