Subramanian Veerappan

Data Scientist Machine learning engineer Data Analyst

Self-driving car built for the Renault DIY Robocar Grand Prix

About

Subramanian

Hello ! I'm Subramanian Veerappan (did you skip my name ? It's okay, you can call me Subra). I'm a full stack data professional with experience in data collection, processing, storage, querying, visualization, prediction and deployment.

Uff, that sounds like a lot right, but yeah, when you work in a startup, it's important to step up to the needs of the organisation and that's how you learn new things. Thanks to my Software Engineering background, which helps me pick up technologies faster.

While I have worked with Machine learing, Data Analytics, Data engineering, & MLOps, my forte lies with Data Science & Analytics

Data + Python + Algorithm = Love ❤️.

Projects

Latest Work

  • Want to export data from Snowflake to G-Sheet ? There are some existing tools like Coefficient, but they are pricy and has some limitation when it comes to efficiency & fetching data from multiple tables. Hence, I developed SnowSheet. SnowSheet helps democratise data by making it available in G-Sheets.

    With SnowSheet, you can write a SQL query as usual in the Snowflake UI, combining multiple tables, conditions, etc. All you need to do is add the link of the worksheet to SnowSheet, along with the target G-Sheet. Based on the frequency specified, SnowSheet excutes the query and writes the resultant data to the target G-Sheet.

  • Magicpot, as the name suggests, it's a collection of disconnected modules, that share the common objective of easing the life of data professionals.

    Have you tried accessing G-Sheets with Python ? It requires tons of libraries to be imported everytime before we can access the G-Sheet. As a data professional, you might have to write the output data from a Jupyter Notebook to a G-Sheet often and you don't want to waste time importing the same libraries over and over again.

    Instead you can use Magicpot, a python package that enables you to read and write data into G-Sheets with just 2 lines of code.

    Another module inside Magicpot is xBridge, as way to share data between different notebooks in Deepnote project

Skills

What I Do

  • Machine learning

    I have successfully implemented ML projects ranging from churn prediction to recommendation systems, with a track record of delivering impactful solutions to complex data-driven challenges.

  • Statistics

    With a strong background in statistics from my engineering background and analytical experience, I've honed my ability to distill complex data into actionable insights, balancing technical expertise with practical application.

  • Python coding

    I have over 9 years of experience working with Python, leveraging it for diverse applications in data analysis, machine learning, and general software development.

  • Data modelling

    I possess advanced data modeling skills using dbt and Snowflake, ensuring efficient transformation of raw data into structured, analytics-ready datasets for actionable insights.

  • Visualization

    I specialize in data visualization using Tableau, transforming complex datasets into intuitive and impactful visual narratives for informed decision-making.

  • Model deployment

    My experience in Software engineering comes handy in architecting MLOps solutions, ensuring streamlined and scalable model deployment workflows

Tools

Tools & proficiency

Analytics & warehousing skills

  • dbt - 0%
  • Snowflake - 0%
  • Tableau - 0%
  • Looker Studio - 0%
  • Python - 0%
  • SQL - 0%
  • Terraform - 0%
  • AWS - 0%

Machine learning skills

  • Sklearn - 0%
  • Keras - 0%
  • BentoML - 0%
  • MLflow - 0%
  • Flask - 0%
  • PyCaret - 0%
  • PyTorch - 0%
  • OpenCV - 0%
Brag doc

Accomplishments

  • Built Snowflake Object Manager package to simplify Snowflake infrastructure management (Febuary,2024)
  • Implemented Dynamic tables in Snowflake for near real time data updates (Febuary,2024)
  • Improved search algorithm by using Wilson score confidence interval for a Bernoulli parameter (January,2024)
  • Built SnowSheet to export data from Snowflake to G-Sheet (January,2024)
  • Took over data engineering and made the first data ingestion into Snowflake (December,2023)
  • Identified and analysed the significance of leakage on the platform, which followed a company wide action + drafting new policies (November,2023)
  • Formulated and build Marketplace KPI tree metrics (October,2023)
  • Reduced merge request pipeline build time to 1/4th by implementing dbt defer (September,2023)
  • Rebranded dbt catalog as xChange catalog (September,2023)
  • Won the first xChange hackathon by building a Recommendation System that recommends similar listings (September,2023). The same project was later productionised and deployed in Kubernetes
  • Developed Magicpot library to help faster adoption of Deepnote (August,2023)
  • Took ownership of Gitlab project & CI/CD pipelines (July,2023)
  • Churn prediction model for Account Management team to predict churn in advance (July,2023)
  • Optimised core models by implementing incremental models and snapshots (January - Febuary,2023)
  • Formulated the SQL style guide for our dbt project (January,2023)
  • Built a data product that gives container market insights and generated quarter to a million in revenue per year (October,2022 - April,2023)
  • Developed company level overview metric dashboard for the Account Management team that helped increase upsells & cross-sells by 14k per month (July - August,2022)
Contact

Get in Touch

To Top