Jul 2022 – Sep 2024 · 2 yrs 2 mos

• Document Understanding at Intuit AI
• Engineered an OCR-free gen-AI recognition model saving millions in annual OCR costs, improving customer experience metrics by 42%
• Experimented with Gemini as a Labeler with immense potential in cost and time savings
• I led the end-to-end development of 11 NLP, Vision and multimodal models in production. Experienced in collaborating with product and data teams
• I have 3 patents in ML from my time at Intuit AI

Apr 2020 – Jun 2022 · 2 yrs 2 mos

• I work to develop end-to-end production-ready ML models for Intuit customers. My work includes multi-modal features to build text and vision based models. As a Data Scientist, I focus on solving the customer problem at hand in the simplest and most efficient way possible while making sure my solutions are explainable and reproducible. I like to be current in my knowledge of AI, NLP and Vision related research as well as contribute to the Data Science community in any way possible.

Jun 2019 – Sep 2019 · 3 mos · San Francisco Bay Area

• Converting cashflow forecasting to a regression problem and bench marking regression models against classical forecasting techniques.
• Developing novel deep learning techniques for improving time series forecasting for Cashflow Data
• Extending the application of Deep Kernel Learning to Cashflow forecasting problems

Sep 2018 – Mar 2020 · 1 yr 6 mos · San Diego

• Artificial Intelligence- Probabilistic Reason and Learning ( Fall 2019)
• Cryptography (Winter 2019)
• Differential Equations (Winter 2019)
• Random Walk Algorithms (Spring 2019)
• Algorithm Design and Analysis (Fall 2018)

Jul 2017 – Jul 2018 · 1 yr · Mumbai Area, India

• Consulted two major banking and non-banking clients throughout my tenure with a focus on Fixed Income and Money Markets. My role involved consulting and software development for the central Web Application with Extensive Back-end development in Java along with some Front-end development in JavaScript.

Jun 2016 – Jun 2017 · 1 yr · New Delhi Area, India

• Used Stanford NLP’s GloVe trained word vectors to normalize Twitter data using mediaWiki dumps data in Python
• Trained a transformation matrix on word vectors of common words in the two corpus using gradient descent to verify the hypothesis that usage differences for words in these two corpus could be captured as a linear transformation

Jun 2015 – May 2017 · 1 yr 11 mos · New Delhi Area, India

• Financial Mathematics, Algorithms, Statistics, Linear Algebra and Differential Equations

Jun 2016 – Aug 2016 · 2 mos · Greater Los Angeles Area

• Mitigating errors and reducing costs of human annotated training data for Google Web Page classifiers using Statistical Inference, Machine Learning and Parallel Processing. Project sponsored and mentored by Google Los Angeles
• Optimal single/multiple rater assignments for rating a set of web-pages training data based on a target Budget
• Synthetic data simulation for ground truth rater assignments and evaluation using Random Forest and SVM classifiers

Dec 2015 – Jan 2016 · 1 mo · Malaysia

• Global Community Development Program
• Designed a Server-Client Android Application for an Old age Home 'Noble Care' at Kuala Lumpur, Malaysia
• Product aimed at language-less urgent notification system to the care staff to serve residents better

May 2015 – Jul 2015 · 2 mos · Gurgaon, India

• Data Analytics and Automation - Built an automated data enrichment system for credit-card merchant database using Visual Basic for Applications and Google and Bing APIs.