The best way of learning anything is by building something.
Competitors Near Fulfillment Centers
Situation I helped a supply chain team during the height of covid to recruit workforce for fulfillment centers, not related to engineering, but to operations like area managers, operations managers, and other warehouse personnel. The usual suspect was to recruit people out of Amazon fulfillment centers. I thought the world didn’t just revolve around Amazon and there must be other companies where to hire from.
Task I wasn’t assigned to build a solution, just to source and hire workers, regardless of the process. I used Google Maps to find distribution centers near a Walmart fulfillment center to see what was in the area. Other than the usual Amazon buildings, I learned that there were big distribution centers from other companies at those locations. However, the process of searching google maps to see what was around a Walmart fulfillment center was too time consuming.
Action I built a dashboard in PowerBI collecting data from Google Cloud API using Python/Pandas to build a dataset of 20,000 data points. The map points to all competitor locations near fulfillment/distribution center. I previewed the dashboard with a leader in supply chain who welcomed the project and asked to preview with their team.
Result The dashboard was shared with dozen of users where I gathered feedback and improved the dataset (still within the free credit).
- PowerBI for data analytics
- Python/Pandas for data engineering
Automating sharing talent with engineering managers
Situation Talent sourcing is like being a headhunter, you have to talk to a lot of prospects, attract passive talent or speak to talent that hasn’t applied to a job at your company. Also partner with hiring managers for specific jobs or to build talent pipelines for a skillset. However, it is complicated to share talent across teams when the available options were manual and too time consuming
Task Another side beta project just to see if I could do it. I wanted to automate the data using reports from the CRM to some sort of dashboard where engineering managers could select candidates that were ‘headhunted’ by the sourcing team.
Action I used Python to collect datasets from the CRM, the data collection was automated using cron jobs, then Pandas to clean the data and separate it to master files. Built a sharepoint page using Power Automate to automate the delivery and daily update of the master files, where the sharepoint page had a way for hiring managers to select candidates and I would get notifications on the backend about any selection.
Result The web app was shared with hundreds of users (hiring managers, recruiters, sourcers) where I captured feedback and streamlined the process of delivering candidates across multiple teams.
At Georgia Tech
- Used Python to collect a dataset 100K+ records from an API to build a node graph. Visualized the relationship of the nodes with a tool called Argo Lite. Used SQLite to do exploratory analysis on the dataset.
- Used OpenRefine to clean a dataset. This is an open source tool similar to Alteryx, but it doesn’t have the drag and drop interface.
- Used Tableau to visualize a dataset. During that time I learned PowerBI and Power Query. Which I used for work projects.
- Used Hadoop, Spark, Pig, and Azure for exploratory data analysis. Used Java for Hadoop and MapReduce, analyzed data with Scala, Spark, and Databricks. Analyze data with AWS using EMR and Pig. Used the Azure ML Studio interface to explore ML algorithms.
- Implemented the page rank algorithm, a random forest classifier, and learned basics of scikit-learn.
- Peer-reviewed hundreds of research papers of classmates in UX research for my class in Human Computer Interaction. Also read about a dozen research papers from conferences.
- End to end design and development of a database including EER, IFD, SQL, and a group project to build a web app dashboard to create reports from a dataset. The web app was built mostly by a classmate. But I am trying to build my own version for my own learning using Node, React, Kubernetes.
- Developed an Android app in Java to encrypt messages. Software development lifecycle in a group project to build another Android app to compare jobs. Used Github for coding collaboration. Built design specs, UML design, test plans, user manual, responsible for some user interfaces. An individual project to write TDD and test cases in Java to build a command line interface to process files.
Other Personal Projects
- Building my blog from scratch with HTML and CSS. Blog design
- Responsive CSS Country Flags. Responsive CSS Flags
- Responsive UI Prototype to find a dog. UI Find Dog
- Old Newspaper Moon Landing with HTML and CSS, intentionally not responsive but exact newspaper dimensions. Old newspaper with only HTML and CSS
- UI Ecommerce Based on Sams Club with Harry Potter theme. Sams Club ecommerce
- After doing the old newspaper more than 6 months ago, I got rusty in web dev and was trying too hard to use inline-block and forgot about flex. I started an exercise to design a classifieds page from an old newspaper (see example here) then it evolved to inspirational quotes. See the result in: Inspirifieds