Bhavish - Machine Learning Pathway

Things Learned
The basics of web scraping using beautiful soup and leant about pickle files. I learnt how to assess the HTTP requests a website makes to get information and to observe the same in the developer options of the web browser. I was introduced to BERT for the first time and my research has led me to understand its application in the Marketing Tech industry for SEO and chatbots.

Highlights

  1. Webs craping (new method previously i used an api to do it automatically)
  2. BERT research requires a new ML pipeline depending on the application

Meetings/ Trainings
I have attended most trainings and meetings

Goals for the upcoming weeks
Would like to implement a simple intent based recommendation engine and next week would like to design an architecture for use case of our project.

Tasks completed
Completed all the tasks in team meetings
I have setup email and slack channel
Having problems with GitHub account setup

Final Assessment:

I got to a know a lot of different people from the Stem Away team. I different set of hands on skills with respect to machine learning and data collection. My experience before stem away did not involve complex data collection and i had never gathered data to solve an NLP problem. I scraped a couple of discourse forms to generate pickle files. Learnt a lot about BERT right down to the math. In my course I had learnt the math of basic algorithm up to neural networks. Using my experience to create a BERT based pipeline require quite a bit of research. Evaluating results was challenging as this time there was no right or wrong answer. I was able to voice my opinions and share my experience with data imbalance and specific metrics like confusion matrices. I also learnt a lot about git and the opensource discourse forums and their theming plugins and mechanism. Brushed up on my HTML, CSS and Javascript in the process also got hands on with ruby. Overall a good experience.