My journey as a computer scientist

About me

I am Sid and I currently do my Master in Data Science at TU Berlin and EPFL. Besides my love for working with data, I am mesmerized by Blockchain. Both - Data Science and Blockchain - have massive potential and will surely shape the future. Being part of this future is my motivation to go beyond my duties as a student and always challenge myself with workshops, hackathons and own (research) projects at a PoC or MVP level. My life goal is to make sure we use data and Blockchain to improve our all’s life. In this blog I document my journey of going for this goal. Feel free to check out how I am doing by looking at my current and past projects as well as my resume here and gitHub: sid030sid.

Current projects

EcoLens

EcoLens is a project initiated by the Zero Emission Group (ZEG) at EPFL. The project’s aim is to create a mobile app for easily finding the most sustainable meal of on any day on the campus of EPFL. While doing so the app will also make use of Gamification in order to persuade its users to actually buy the most sustainable meal of the day. I generally joined the student association ZEG in September 2022 because all its members are students who also want to use their knowledge and skills for the better. It is always awesome to work with like-minded people and I am happy to contribute in the EcoLens project as a SCRUM Master and Full-Stack App Developer. Regarding the app development my main responsibility is the gamification layer of the app which I will code in Swift and React Native. I wanted to do this part so that I can make use of my knowledge in Gamification obtained in the course “Digital Service Engineering” at TU Berlin. Besides my coding and organizational tasks, I will also propose to the Data Analytics Lab of EPFL to make a research project about the impact of EcoLens on the consumption at EPFL with me as the researching student. All labs at EPFL actually initiate such semester projects themselves so I am hoping my proposal will be accepted as one. Since the app is not launched yet, the research would be about the current state of the consumption on campus as well as the methodology the app should follow in order to make a statistically sound assessment on the impact of EcoLens. It would thus be a collection of several ways to capture metrics for the impact and their respective pros and cons. The best one will then be chosen and implemented in the EcoLens app.

Running foundations as a DAO

During my exchange at EPFL, I meet very inspirational and motivated people. One of them is running a foundation in Sweden called “Moderskeppet”. They are doing diverse projects in the field of environmentalism and social justice. Due to they strong belief in Blockchain they want to run their foundation as a DAO. Since I have always wanted to challenge myself to do a DAO project in order to better grasp the concept and extend my Blockchain developer portfolio, I offered my Solidity skills to create a MVP solution for them. After some talk, it is now clear that I will also teach the SCRUM method for a better collaboration due to the agile nature of the project. Although we only make use of the product backlog and the roles: product owner (= the foundation) and SCRUM master/team (both me), it already helped the project a lot. Within this project I will also code a mobile app using React Native for interacting with the DAO’s smart contracts. To be fast, I taught the foundation how to visualize their DAO app in Figma and do wireframes in Miro. So far, I have used my semester break to code the DAO’s governance token and governor contract. The later manages proposals and votes. Its code is based on Openzeppelin Wizard’s. The next step is unit testing all smart contracts using Hardhat. To properly end this project, I plan to document the project in a paper and hand it in at the LBIS conference as a submission for “Administrative Sciences”.

Spatio-temporal clustering for detecting epidemics

During the current semester I will assist intelligent Global Health - a special research group of the Machine Learning and Optimization Labrotory at EPFL for developing AI to improve healthcare in resource-limited settings. For the DYNAMIC project I will use real-life medical big data from Rwanda and Tanzania to find clusters of patients’ demographic and symptoms. The found clusters will subsequently be analyzed regarding time and location which gives insight about potential epidemics. In pursue of doing this, I plan to apply and compare various clustering methods, i. a. K-Means, Gaussian Mixture Models and Agglomerative Hierarchical clustering. I am happy to be part of this project due to its meaningfulness and what it will teach me. Besides applying unsupervised learning, I will definitely gain more practical data science experience in the medical field which requires a higher focus on validating machine learning models since people’s life are at stake. In the end of this project I will produce a data pipeline for detecting epidemics as well as as an academic paper about my whole work.

Past projects

Bimcards

Bimcards is a web app a colleague and I built on behalf of “planen-bauen 4.0 GmbH”. Since this was my first web development project ever, I started it with nearly no knowledge (apart from the coding skills obtained in uni and on the side). As a consequence of this project, I became a Full-Stack Web Developer. To build the app, we used the MERN stack and Kanban as an agile work technique to handle all the consistently changing requirements of the owner of the app. I am very thankful for having had the opportunity to work on this project as I learned so much from it. My main responsibilities were frontend development and database design. For the latter I additionally did courses in building and managing MongoDB’s NoSQL databases. Offered by MongoDB Uni these course particularly taught me about data modeling and performance optimization.

unbiasedTwin

With the aim to make it easier for companies to do anonymous job applications, I came up with the web app called unbiasedTwin. In a nutshell, unbiasedTwin lets companies create upload portals people can use to submit their job application anonymously. This project challenged me throughout the entire life cycle of a web app and required me to be an entrepreneur and developer. I built the app using the MERN stack and deployed it using Heroku’s PaaS. However, formulating and properly capturing the idea of unbiasedTwin was the hardest because I just wanted to code right away. Therefore, my key takeaway from this project is the necessity of only starting to code once everything is decided on at a conceptional level. To achieve this, I initially created user diagrams and wireframes in Miro. Regarding coding, the tricky part of the app was the built in chat as well as the automatic Email messaging system based on Nodemailer. Another big lesson I learned from unbiasedTwin is that one should not underestimate the maintenance of an app. Even though using cloud services like Heroku and MongoDB, maintenance still required a lot of time and energy which is why the app is currently down.