About Me

How did I end up in Data Science?

Given the remote location of my undergraduate campus, I found myself relying heavily on online shopping. My curiosity about product recommendations sparked a fascination with Analytics. Through a blend of e-learning, flexible courses, and internships, I pursued knowledge in this field. Drawn by the excitement of data science and machine learning, I was inspired to take on a new challenge – dipping my toes into the waters of software engineering. (pun intended: I have majored in Ocean Engineering)

Upon completing my degree, I joined Barclays, where I immersed myself in development role. Simultaneously, I dedicated time to self-initiated machine learning projects, nurturing my passion. However, my aspiration for a full-time career in Data Science led me to join Kotak Securities. Here, I engaged deeply in customer and marketing analytics, thriving under an exceptional manager.

Yet, a persistent concern lingered – the absence of formal education in Analytics and Machine Learning. Fueled by a lifelong dream of pursuing education abroad, I turned to my manager for guidance. His invaluable advice steered me towards the Business Analytics program that blended my robust technical acumen with a strong business focus. This journey proved transformative, refining my ability to tackle complex problems and craft effective analytical solutions.

More resolute than ever, I am now eager to embrace full-time opportunities in Data Science and Analytics.

Skills

  • Languages: Python | MySQL | R
  • Database: Mongodb | Big Query
  • Frameworks: Sklearn | PySpark | Tensorflow
  • Tools: AWS | Google Cloud
    •

    Specialities

  • Exploratory Data Analytics
  • Experimentation & Analysis
  • Supervised & Unsupervised Machine Learning
  • Data Storytelling
    •

    Resume

    Experience

    • Data Scientist, Practicum Project | Qualcomm | San Francisco, Remote

      2022 — 2023

      As part of the MSBA, identifying user needs and preferences in the mobile device markets to provide business insights and improve branding. Conducted market research and survey to better understand the competitor landscape and consumer preferences.

    • Associate Vice President | Kotak Securities | India

      2021 — 2022

      Built MMM, propensity data model & deep learning-based segmentation model to increase the adoption of various products. With a strong focus on customer, product, and competitor analytics. .

    • Graduate Analyst | Barclays UK | India

      2019 — 2021

      I honed my skills in software development and data analysis. I worked on projects analysisg customer spend behaviour and fraud detection models.

    Internships

    • Research Assistant | Cambride Judge Business School | Remote

      2018 - 2018

      Scraped pdfs, their title, and publishing year from 215 websites, using Beautiful Soup library.

    • Data Scientist Intern | Kotak Securities | India

      2018 — 2018

      Identified factors influencing the trading Behaviour of ∼3M clients and leveraged them to increase activation.

    • Machine Learning Intern | REdX Innovation Lab | India

      2017 — 2017

      Identified personality traits using tweets, scrapped through Twitter API, and processed using NLTK Toolkit to generate non-traditional credit score.

    • Data Scientist Intern | Flytxt | India

      2016 — 2016

      Classified preference of the sim as either primary or secondary for multi-sim users using a random forest model.

    Recent Projects

    Pharma Scrape and Analysis

    One of the most prominent issues during the COVID-19 pandemic was the unavailability of medicines. A few suppliers dominate the pharmaceutical industry, and patients often pay more for medicines due to lack of knowledge. These models would help pharmacies improve inventory management and pricing strategies.

    • Python
    • NLP
    • Word2Vec
    • Beautiful Soup
    • Pyomo

    Text-To-Image Generation

    When in SF, gotta try Generative AI! Fine-tune a stable diffusion model to generate flower images using text prompts. The data was transformed for a CLIP model and uploaded on hugginngface. The updated weights after fine-tuning are also uploaded on huggingface.

    • Stable Diffusion Model
    • Tensorflow
    • Hugging Face
    • Keras

    Spend Behaviour & Social Network Analysis

    Evaluated how the percentage of textual description v/s emojis used to describe transactions affect customer lifetime. Calculated Social Network metrics- clustering coefficient (triplets), page rank at each point in a lifetime. Predicted customer transactions using RFM framework, social network metrics.

    • PySpark
    • GraphFrames
    • RFM
    • MLib

    Other Projects

    sq-sample26

    Locate Pizzeria in San Francisco

    Scrapes resturant information, their geolocation and saves in MongoDB
    sq-sample26

    Who are the top contributors of a Repo?

    Extracts the information of top contributors on a github repository using github api and stores the data into mysql database.
    sq-sample26

    Gift cards sell above face value?

    What percentage of Amazon gift cards sells above face value?
    sq-sample26

    K-means clustering & PCA

    Algorithms to deal with high dimensional data. As well as affect of initial clusters and the change in clusters as algorithm iterates to get optimum cluster centers.
    sq-sample26

    K-Nearest Neighbor Classifier

    Classifies iris flowers and comprehend the changes in performance of k-nearest neighbors as k takes on the different values.
    sq-sample26

    Treatment Effect

    Assigning a Different Hotel Room: How it Affects Cancellation?
    sq-sample26

    Regression Discontinuity Design

    Quazi Experiments: Establish causality and draw reliable conclusions from observational data.
    sq-sample26

    Predict Purchase

    Implements logistic regression using stats package in Python. Considers the interaction between covariates.Illustrates different ways variables can interact.