Day 1, Getting Started
An Introduction
We are a group of four students, Steve, Brenda, Jules and Sureka, at London’s leading software development bootcamp. We’ve been given two weeks at the end of our 12 week course to be set loose on anything in the world of tech. We were grouped together under our shared common interest in machine learning.
Deciding the Project
We sat down for an hour and hashed out what in the wide world of machining learning would be interesting for us to do and achievable by novices in both machine learning and python our chosen language. We narrowed it down between a game, connect 4, and natural language processing, to identify hate speech in tweets. We settled on the game as it has easily defined parameters to define success, you win or you lose. We also thought that it would make a cool demo for people to play with in 2 weeks time. We later found out another team is also had the same idea. We decided to carry on regardless and look forward to AI-bowl at the end of 2 weeks.
Creating the MVP
Next up on the to do list was define what are minimal viable product was for this project. This will help us keep focussed to get something out that we can then build on for success. Our MVP is a computer can play connect 4. We broke this down into user stories and promptly added them to our Trello board, the project management web app. During this discussion we also settled on our team name, Deep Red&Yellow. A play on deep blue, the famous chess AI and the red and yellow counters that are used in the game.
Setting Expectations
In any team project it’s important to get the team aligned and iron out any potential problems at the start. We tried to do this by setting out what we expect from the team as well as some of our goals. We all agreed that we were going to work as an agile team. Working in sprints, having stand-ups and retros to keep everyone on the same page and constantly improving. We also wanted to focus on documenting our learning and process. Hence, this blog was born.
Getting Started
First things first, re-search the tech stack. We went for tensor flow as our machine learning library. We had seen that it is not only one of the most popular but also mentioned often in job postings. After all jobs are weighing heavily on all our minds now so close to the end of our time on the course.
During the course we made extensive use of test driven development, TDD. Although somewhat foreign to us when first starting the course we have now grown to love and appreciate it. We are happy to explore testing in a new language, confident the course has given us the skills to do so.
Hello TensorFlow
Whats the way all programers get to grips with something new. They start simple, they start hello world. This is exactly what we did, taking our first small steps into an extensive library. We checked out a small blog that explained what tensors and flow actually mean as well as how to use them simply. It’s a bit daunting going from addition to machine learning. I’m optimistic we can rise to the challenge within a week.
Steve Kahn
Subscribe for more updates from Deep Red&Yellow.
