• Spotify Wrapped 2022 logo with hashtag #SPOTIFYWRAPPED on purple background.

    Spotify 2022 Wrapped

    I was the Insights Lead for 2022 Wrapped, managing a team of Data Scientists and User Researchers.

  • Player first Spotify homepage

    Owned multiple experiments in partnership with engineering to validate the hypothesis that Spotify web visitors are better served if shown the web player instead of the existing homepage. This change added 2 million Monthly Active Users.

  • Screenshot of a data table featuring columns such as notes, batch_transaction_count_match, result, amount, organization, provider, created at, status, and approved by. Some rows contain notes like 'Approve' and 'Abnormal Payouts Count,' with results marked as 1 or 0. Other rows show amounts, provider details, timestamps, statuses, and names in approved by column.

    Uber driver payment anomaly detection

    Uber disburses payments to millions of partners (drivers and eats merchant restaurants) every week in more than 60 countries. I built a decision engine to identify payments that are anomalous and worked with the engineers to implement a fully functioning tool that stops anomalous payment going out automatically.

  • Line graph comparing pre and post implementation rates, showing a 2% increase pre-implementation and a 4% increase post-implementation. Includes text about methodology and market group onboarding event.

    Measuring Financial Impact of Promotion Manager

    Without a proper A/B testing setup available due to retail industry operational complexities, proving the financial impact of using Rubikloud's Promotion Manager product was challenging. I designed, formulated, implemented and educated the analytics and sales teams the methodology behind estimating the financial impact. The code base is used to drive demo, pilot and subscription renewal agreements.

  • Python code snippet for bootstrap mean estimation using NumPy.

    Standardized Campaign Measurement Module

    Developed Rubikloud's standardized campaign measurement module that was deployed across all clients for Customer Lifecycle Manager product

  • A group of people wearing purple shirts sitting around a conference table with laptops, engaged in discussion. A whiteboard and a sign with 'RAINBOW RAILROAD' and 'Data Philanthropy' are visible in the background. The setting appears to be a collaborative meeting or workshop.

    Aimia Data Philanthropy with Rainbow Railroad

    Led a team of 13 in Aimia's 2018 Data Philanthropy event. Built a data cleaning and segmentation tool using R, and lead other 5 objective streams to deliver a full suite of operational improvement solutions to Rainbow Railroad, a Canadian charitable organization.

  • Map with red dots indicating locations, black lines for roads, and a dark area representing water.

    Toronto Traffic Jam Hackathon 2015

    Measured the variance of speeds from transit data to map the unreliability of Toronto's road network. This insight would allow commuters to make more informed route choices, and city planners plan events/constructions accordingly. The team won prize money of $5000.

  • Person giving a presentation on stage with an audience, wearing a blue shirt and gray pants.

    2015 University of Waterloo 3MT finalist: Tawsif Khan

    I was the winner of the 2015 Faculty of Mathematics Three Minute Thesis (3MT) heat. In 3MT you are given 3 minutes to explain your thesis to a general audience.

  • Diagram of a one-layer shallow water model showing a wave profile with variable parameters η(x,t) and v(x,t), over a baseline height H. Axes labeled z (vertical) and x (horizontal). Water flow indicated by arrows. Mathematical integrals shown below.

    Optimal Sensor Location Research

    For my master's thesis work at the University of Waterloo, I worked on the optimal sensor location problem in the Infinite Dimensional Systems research group. My work specifically focused on finding the optimal discretization technique to solve the Navier -Stoke’s equations. I presented this work at the American Control Conference in 2015.