Nazanin is a Senior Data Scientist at Intuit. She and her team harness the power of AI to turn typically tedious financial tasks into simplified customer experiences.
How would you describe your role to another engineer?
NZ: As a data scientist at Intuit, I deal with Intuit customer data and devise AI models that can digest and process data and produce informative outputs that can improve customer’s experiences and provide them benefits. As an example, the automated receipt extraction service processes thousands of receipts and extract informative information in a couple of minutes.
What is the most rewarding part of your role? When do you feel most fulfilled?
NZ: When I see the models that I built are actually helping our customers and lead to revenue and growth of my company. In addition, when I see that I am able to solve a problem with new methods that can lead to patent or presented internally/externally.
What made you choose Intuit and why were you drawn to the company?
NZ: I heard great things about intuit products and used TurboTax and mint before. I saw it as a growing company as credit karma also joined. It has a great future ahead for applying ML/AI.
The AI model you used leverages deep learning, computer visions, and neural networks – what role do each of those play in this new product innovation?
NZ: Deep neural networks based on transformers are used for NLP task which is name entity extraction (NER) from the texts and is the state of the art model in this domain. Computer vision techniques were used to detect the structure and template of the documents which is also based on new findings in 2019 and 2020 pieces of research.
Why is it so important to eliminate typically tedious financial tasks and save time for small business customers?
NZ: From accountants’ side, they provide accounting service for many customers and reconciliation of the documents is the most time consuming part for them. This automation aims to decrease this time for them which leads to a better product experience.
What do you see as the next iteration of this initiative?
NZ: Expand this automation to more geographies (support new languages receipt extraction), and the end to end automation of the whole document reconciliation.
What are you most excited for in 2021?
NZ: I am actually very excited about a new project that I joined in lending space. I think it is a great domain for Intuit to apply AI/ML and can revolutionize the whole lending space.