Skills and Expertises

Along the journey in my career, I have worked on various projects, all of them may not be worth of mentioning. I have handpicked few of them here along with the tools and methodologies used.

Academics

  • Digitization of class notes, preparation of project reports
    • Tools
      • \(\LaTeX\) as the rendering engine
      • R with Sweave and knitr packages
  • Performing computations for ‘Design of Experiments’ problems
    • Tools
      • C programming

BCausE Enterprise Pvt. Limited (2015-2016)

  • Predicting the names of human body parts from a corpus of digitized prescriptions using text classification
    • Tools
      • Heavily relied on regex for data cleansing
      • NLP libraries based on Python
      • MongoDB at the back-end
      • Flask for REST API as front-end

IBM India Pvt. Ltd. (2016-2021)

  • Analysing customer complaints (a.k.a tickets) from various sources and predicting root causes using clustering and text classification respectively
    • Tools
      • Python (as machine learning base)
      • SQL at the back-end
      • Little bit of ShellScript to integrate the model with other modules
      • Flask for REST API as front-end
      • docker for packaging the ml application
      • Kubernetes used for deployment of the docker image
  • Serving search queries to customers by designing a small scale search engine built on customer Feedbacks and FAQs using information retrieval algorithm
    • Tools
      • Singular value decomposition to understand word-to-word relationship
      • Pandas as data processing toolkit
      • Numpy for matrix facorization implementation
  • Understanding the effects of various factors like advertisement, promotion, weather on sales of several commodities, analysing ROI year-on-year basis and forecasting the furure sales using Marketing Mix Modeling
    • Tools
      • R with dplyr as the data processing toolkit
      • Base R for model development
      • SQL at the back-end

IHS Markit and S&P Global Inc. (2021-Present)

  • Analysing shipping data on a daily basis
    • Tools
      • PySpark as the data processing engine
      • Excel for adhoc analysis and reporting
      • AWS as the cloud platform
  • Cargo grade identification using text classification
    • Tools
      • Python (as machine learning base)
      • S3 for Local Deployment
      • ShellScript for automation
  • Building data pipeline incorporate external data from various sources
    • Tools
      • R for developing data pipelines
      • ShellScript for integration as automation
      • AWS as cloud platform