Internship Description

Arusha Kelkar, ’20SEAS, interned with Tamer Fund for Social Ventures portfolio member Plentify, a sustainable energy startup in Cape Town that aims to enable South Africans to switch to reliable and sustainable energy sources while minimizing utility bills by harnessing the power of AI and advanced grid technology. Arusha was a data science intern, predicting electricity demand using information from various device usages, working with data from water heater IOT controllers and building machine-learning models for enhancing efficiency of electricity usage.

Plentify is a Cape Town-based startup providing reliable and sustainable energy and water services. I worked on two projects throughout the internship: parameter estimation for electric water heaters, specifically the heating rate and thermal coefficient; and implementing thermal models on the Plentify fleet of geysers. The parameters (heating rate and thermal coefficient) estimated are used in calculation of dispersion score, which is a measure of how much more energy an electric water heater (geyser) needs in order to satisfy a user’s future needs relative to how much energy the element (thermostat) is physically able to apply in that time, and determines how urgent it is to turn on that geyser’s element. Thermal models should be able to accurately estimate the amount of usable energy that can be extracted from the geyser to match the hot water supply to demand, which is of utmost importance for user satisfaction.

I worked on accessing data from the Plentify database using SQL queries. I also worked on data manipulation and data cleaning in Python using various Python libraries like Pandas and Numpy. I particularly did a lot of data analysis using methods to apply functions on each row of the Pandas dataframe and speed up the execution of the same. Use of Python libraries like matplotlib and seaborn was an integral part of creating data visualizations for detecting outlier conditions in the parameter calculations throughout the internship. Understanding of object-oriented programming in Python was an essential part of implementing the thermal models on the Plentify fleet. Everything that I learned at Columbia University was extremely helpful for tackling technical issues and approaching various data problems effectively and solving them efficiently.

My internship was remote, which posed certain challenges. Regular meetings with the team along with interesting problems to solve everyday kept up the momentum of the internship. Sharing of personal experiences on issues helped overcome the challenge of lack of personal contact. Understanding the working of the electric water heater posed some challenges initially but these challenges made the experience more engaging and memorable.

I learned a lot about how startups navigate building a product from scratch and what keeps them motivated. Throughout the internship, there was a sense of contribution towards the final product, which was extremely satisfying. The interaction within the various teams of the organization made the entire experience enjoyable. Being able to express the problems effectively and reaching out to various teams for feedback was something valuable that I learned. The entire process made me more aware of the years of hard work it takes to build something user-friendly and including innovative ideas that would engage the customers effectively. It has reinstated for me the sentiment that patience is the key to success.