Technical Areas Involved In/ Covered:
Natural Language Processing Algorithms
- Bag of Words
- TF-IDF Vectorization
- BERT, DistilBERT
Machine Learning Classification Models
- Deep Neural Network
- Decision Trees
- Random Forest Classifier
- Support Vector Machine (SVM)
- Logistic Regression
Agile (Scrum) Product Development
- Python - Numpy, NLTK, Gensim, Scikit Learn, Pandas, Keras, PyTorch, Beautiful Soup, Selenium, Scrapy
- Google Colaboratory
Soft Skills Developed and Applied
- Research: Before each session, I thoroughly researched up the right material and the best way to deliver the content as most of the participants were fairly new to machine learning. I provided them with sufficient material to help them understand the basics and if interested dig deeper with additional resources.
- Mentoring: This being my first experience in mentoring a team of students who were fairly new to the field of machine learning. I wanted to help them hone their basics and no be afraid of jumping into the mathematics behind the algorithms. To achieve this I inculcated a sense of curiosity and admiration for the field. Towards the end of session I was able to see the students becoming interested in research upon novel algorithms. They were more willing to go the extra mile and try out something new. The best example of this was them stepping up to develop a novel hierarchical classification model that was out of the scope of the session.
- Project management: I guided and mentored the students individually and in a group to help them understand the task, ensure they were always felt included and comfortable, and further supported them in completing the tasks both technically and morally. For better management of the project, I conducted daily stand-ups, bi-weekly sprint retrospectives, and occasional sprint planning sessions.
- Teaching: I conducted tutorials and code walkthroughs to help the participants understand various concepts and develop basics in topics ranging from Web Scraping to Natural Language Processing. Along with these tutorial sessions I also researched on easy to understand the material which I shared with the participants.
- Team Building: The participants were graduate and undergraduate students from different parts of the world. It was challenging to help them open up to each other and work in a collaborative manner. Looking at this I conducted various team building exercises in the first week which included but was not limited to the “Goal Corridor”. This helped them understand what they and their peers were expecting out of the experience and how common their goals were. When the participants were comfortable with each other I divided them into groups to further enhance the collaboration. What started as a group of strangers became a cohesive team that was willing to support each other.
- Collaboration: I collaborated seamlessly with participants who were all located on different time zones. This allowed me to gain experience in virtual collaboration.
- Developed recommendation models using data scraped from Atom Forum. We compared the performance of various NLP algorithms - BERT, Doc2Vec, TF-IDF, BoW - to find the best-performing algorithm.
- Developed a novel hierarchical classification model to classify any query topic to the forum as well as the category within that forum that it belongs to.
Meetings Attended/ Conducted
- Team and team lead Meetings
- Tutorials, Workshops, and Code walkthroughs
- Sprint Planning and Retrospective
- ML Weekly Lead Meetings
- ML Training Sessions with Colin
Tasks as a Lead
- Researching and preparing lectures and workshops.
- Setting up Github and regularly maintaining the repository.
- Delivering lectures either twice or thrice a week.
- Preparing assignments for participants.
- Tracking participants’ progress through weekly tasks and deadlines.
- Providing feedback on work done.
- Setting up and sharing reading material on Notion for the participants.