I do not just look at data, I interrogate it. Every dataset has a story buried under the noise, and I am the kind of person who keeps asking what happened, why it happened, and when did the pattern shift until the numbers actually make sense.
With a Masters in Data Science from Stevens Institute of Technology and a Computer Science background from Amity University Mumbai, I have built the toolkit to back up that curiosity, from machine learning and statistical modeling to cloud infrastructure on AWS.
Data-driven cross-selling strategies that contributed to over $10 million in revenue at Motilal Oswal, predictive models that cut data errors by 20%, and algorithmic approaches that improved trading decision accuracy.
I have worked across financial services, edtech, and AI startups, and in every role, the thread is the same: dig deeper, question the assumptions, and let the data lead the decision.
I hold certifications in Bloomberg Market Concepts, AWS Machine Learning, and Agile Project Management, and I have published research in peer-reviewed journals.
When I am not drilling into datasets, you will probably find me tracking flight routes on Flightradar24, reading up on the latest geopolitics, or keeping tabs on what is moving in the financial capital markets.
Uplifty AI · Austin, Texas, USA · Remote · Full-time
Tenure: Sept 2025 – Present
• Built and tested ML prototypes to explore personalization features, improving early user engagement by 12%.
• Worked with product managers to convert business needs into technical roadmaps for AI-driven solutions.
• Researched AI adoption trends and competitor tools, guiding product positioning and feature priorities.
• Analyzed user behavior data to suggest feature updates, enabling data-backed product decisions.
Stevens Institute of Technology · Hoboken, New Jersey, USA · Part-time
Tenure: Oct 2024 – May 2025
• Analyzed admission data using advanced statistical techniques and problem-solving approaches, providing actionable insights to the Department Chair of Systems and Enterprises for informed decision-making.
• Developed interactive dashboards to manage faculty load across multiple semesters, leveraging Python for data cleaning, integration, and visualization, streamlining resource allocation and planning.
• Managed inventory management solutions, increasing efficiency and accuracy in managing departmental assets by 25%.
• Provided comprehensive technical support to faculty members, technology troubleshooting needs, enhancing teaching and research productivity.
• Optimized administrative workflows, integrating technical solutions to reduce manual effort, ensuring timely completion of critical tasks, and improving departmental operations.
Stevens Institute of Technology · Hoboken, New Jersey, USA · Part-time
Tenure: August 2024 – May 2025
• Oversaw a community of 400 members on Ducklink, facilitating engagement and communication through regular updates and interactive events.
• Organizing and leading technical workshops and seminars annually, increasing member participation target by 40% year-over-year.
• Boosted membership by 30% over a 12-month period, expanding the association's reach and engagement within the graduate community.
Stevens Institute of Technology · Hoboken, New Jersey, USA · Part-time
Tenure: September 2024 – December 2024
• Evaluated and graded approximately 100 student assignments and exams per semester, maintaining a grading accuracy rate of 95%.
• Provided detailed feedback on 50+ assignments per semester, improving student performance and comprehension by 20% based on survey results.
• Managed an average of 5 hours per week for grading and administrative tasks, ensuring timely completion of all grading responsibilities within established deadlines.
Stevens Institute of Technology · Hoboken, New Jersey, USA · Part-time
Tenure: June 2024 – January 2025
• Led engagement programs, Pre-Orientation, and Orientation Week to support new students in their college transition. Offered continuous guidance throughout their first year to ensure a successful and seamless adjustment.
• Assisted new students through the course selection process, ensuring they chose courses that matched their academic aspirations. Linked them with campus resources to enrich their college experience.
• Fostered a welcoming community by organizing social and academic events, promoting peer connections, and encouraging student involvement to support their personal and professional growth.
Motilal Oswal Financial Services Ltd · Mumbai, Maharashtra, India · Full-time
Tenure: August 2022 – June 2023
• Played a pivotal role in driving the company's revenue growth through the successful implementation of data-driven strategies for cross-selling multiple financial products, including Mutual Funds (MF), Alternative Investment Funds (AIF), Insurance, and Portfolio Management Services (PMS). During my tenure, the company achieved a remarkable milestone, generating over $10 million in revenue from these products, attributable to my data-driven contributions.
Key Roles and Responsibilities
Kotak Securities · Mumbai, Maharashtra, India · Internship
Tenure: June 2021 – July 2021
• Analyzed intraday and historical equity data to assess trading behavior, spot patterns, and validate price movements.
• Built and tested LSTM and algorithmic models to detect unusual or irregular market activity.
• Tuned model thresholds to reduce false signals and strengthen trading decision support.
• Researched data quality and volume impacts on predictive accuracy to improve market-monitoring reliability.
Hoboken, New Jersey, USA
September 2023 – May 2025
Master of Science (M.Sc.) in Data Science
Provost Scholarship: Merit based scholarship of 10,000 USD.
CGPA: 3.9/4.0
Relevant Coursework:
Probability Theory Statistical Models Applied Machine Learning Optimization Models Deep Learning Data Visualization Marketing Analytics Supply Chain Logistics Time Series Analysis
Mumbai, Maharashtra, India
August 2018 – May 2022
Bachelor of Technology (B.Tech.) in Computer Science Engineering, Minors in Business Management
CGPA: 3.84/4.0
Relevant Coursework:
Cloud Computing Database Management Systems Software Engineering Artificial Intelligence Financial Management Entrepreneurship Development
Developed ARIMA(2,0,2) and SARIMA(2,0,2)(0,1,1)[12] models to forecast Hawaiian Airlines monthly departure delays (non-seasonal) and Atlantic storm frequencies (seasonal), respectively. Executed full time series modeling workflow: ADF tests for stationarity, ACF/PACF for order selection, and AIC/BIC-based grid search for model optimization. Validated model assumptions through residual diagnostics (Shapiro-Wilk, Ljung-Box), confirming white noise and robustness of forecasts across 6–12 month horizons.
Utilized historical data on inflation, unemployment, and bond yields, spanning 844 monthly data points. Preprocessed data to achieve stationarity and ensure model reliability. VARMA model delivered outstanding results with an RMSE of 0.15 and MAE of 0.06. VAR and SARIMAX also performed well, while LSTM faced challenges due to dataset-specific complexities. Bond yields had the strongest correlation (0.86) with the Federal Funds Rate, followed by inflation (0.71). Demonstrated the effectiveness of traditional econometric models over neural networks for this task. This enables stakeholders in finance, real estate, and policy-making to anticipate interest rate trends and make informed decisions. Potential to develop hybrid models combining traditional econometrics and machine learning for even greater accuracy.
The 2024 EPA Vehicle Fuel Economy dataset provides detailed information on vehicle fuel efficiency, carbon emissions, and related attributes. This study uses statistical methods to analyze the data, aiming to identify factors affecting fuel economy and emissions, and to propose strategies for improved efficiency. The study includes data description, pre-processing, descriptive statistics, and inferential analysis, concluding with findings, implications, and future research directions. This analysis offers valuable insights for stakeholders on the environmental impact of vehicles, supporting informed decision-making and sustainable practices.
I Developed an AI model for employee attrition prediction using logistic regression, XGBoost, random forest, and cosine similarity, achieving 83.81% accuracy. It reduced attrition-related costs by 20% and improved retention rates by 15% through this innovative methodology. In addition to this it also provided the HR department with actionable insights to identify potential churners and implement proactive retention strategies and emphasized cost reduction by focusing on retaining existing employees instead of new hires.
We developed a movie recommendation system that suggests movies based on user input, providing suggestions and similarity scores. The process involved understanding the problem and key concepts, preprocessing data, visualizing it, and building the model. We used algorithms such as K-means Clustering, SVD (Singular Value Decomposition), and inbuilt algorithms from the Light FM library. The recommendation system was deployed using the TMDB API. This system includes features like allowing users to watch movie trailers directly on YouTube and displaying trending movies in the recommendations.
I have been working on finding ways to detect and categorize plant diseases using images of their leaves. Sadly, there still aren't any reliable methods available commercially for identifying these illnesses. In my study, I decided to try out two different types of convolutional neural network models – VGG and ResNet. I used a dataset of 61,486 images from Plant Village to train and test my models. These images were taken in labs and covered fourteen different kinds of leaves across thirty-nine categories. The best accuracy rates with the ResNet and VGG models were 81.6% and 70.4%, respectively, for non-inherited indices. I decided to split our dataset into train, test, and validation sets, with 36,584 for training, 15,679 for validation, and the rest for testing. What's interesting is that the deep-learning model using the VGG architecture needed less time to train on colored images compared to other methods, showing potential for efficient disease detection in plants.
I leveraged predictive analytics tools that utilized various models and algorithms for a wide range of applications. Choosing the right predictive modeling techniques was crucial for maximizing the benefits of predictive analytics and making data-driven decisions. To implement this, I acquired the dataset, imported necessary libraries, and loaded the dataset. Next, I identified and handled missing values, encoded categorical data, performed feature scaling, and split the data into training and testing sets. Using a linear regression model, I aimed to predict outcomes and identify key factors influencing the results. In the future, I plan to deploy the model using Flask on platforms like Heroku or GitHub for real-time analytics and wider accessibility.