Collaborating with people from different backgrounds and skills made me thought out of the box and achieve my ambitions.
Nearly a year back, when I was introduced to Python programming during my Bachelors in Computer Engineering, the practical applications it grasps fascinated me a lot. After studying the language thoroughly I started reading blogs and articles on Machine Learning and Natural Language Processing to understand the topics briefly and thus began my journey with online education. Also, I am still exploring novel theories and innovations that take place each day to enhance my knowledge and become a better individual!
“One hour per day of study in your chosen field is all it takes. One hour per day of study will put you at the top of your field within three years”
After exploring the copious amount of resources available on Google, I stumbled upon the Machine Learning course offered by renowned Stanford University. After getting familiar with the concepts used in this course, I started with the Deep Learning Specialization course introduced by Andrew Ng on the Coursera platform. Although I was truly enjoying these well-curated courses, deep down I suddenly felt the urge to work on a real-world project and benefit the community.
After completing these courses, I was reading blogs during my leisure time and I came across the blogs published by Omdena, and then I realized that this is the opportunity I was looking for. I still remember the day I applied as I was nervous about working for a large firm with a diverse community and if I could work according to the requirements. But a few days later, I received news that they had selected me to be part of a 50-member team that would work together for Improving Cross-Border Child Protection over the next two months and my excitement knew no bounds. I am very glad that I took the opportunity that landed me with a community that is so supportive.
“If a window of opportunity appears, don’t pull down the shade.”
We were all invited to the Slack channel where we introduced ourselves briefly before the kick-off call. After our discussion, we came to know how difficult it would be to engage with each other considering the different time zones and backgrounds. Everyone felt the same way in the beginning but as we started engaging more and more by stepping out of our comfort zone, the difficulties flew away just like after working a lot on joining different pieces of a puzzle it gets easily solved and it is rightly said the more the merrier!
In our first meeting with our parent organization, we got a concise idea of what the deliverables should look like so we spent a lot of time researching our problem definition and discussing the possible deliverables that we can develop in the stipulated timeframe. As we all come from different technical backgrounds and diverse skills, we got a bunch of opinions and approaches on how to work.
The approach followed by Omdena that I really admire is the Bottom-up collaboration. This approach gives voice and decision-making power to all data scientists that are working on the problem, keeping aside the top-down approach that relies heavily on the higher-level authorities that decide the workflow and pass it to lower-level employees.
First and foremost, to work in any online platform it is necessary to engage ourselves to do our best so there comes — self-motivation. To continuously take part in group conversations by proposing ideas and innovative approaches it is vital to be self-motivated and work each and every day doing lots of research. Without proper engagement, it is not possible to understand what the other person is thinking or talking about.
Comfort in Learning and Using Digital Tools– The pandemic has taught us the importance of working online from anywhere at any time interval which is the best skill any corporate recruiter is looking for nowadays. We mainly worked with tools such as Slack, Zoom, Google Colaboratory, Drive, Documents, and Slides. Working with diverse time on a completely online platform taught me that everything is possible with diligence to work.
Attaining consensus- It is very important to respect the opinions of the entire team whenever working in a collaborative environment. During our project execution, we investigated all the different approaches proposed by the team and then arrived at the most viable solution that can be delivered to the parent organization.
Knowledge sharing and note-taking– Without telling the team what you think, they can never understand what approach is coming out of your head. So during our meetings and calls, we mainly focused on the ideas and implementation part by understanding everyone’s opinion. This really helped me to enhance my communication skills and burst the bubble of shyness. Besides this, taking notes of opinions shared by everyone in the group helped me to learn a few new techniques and approaches which I never thought would even exist!
Support – This is something that holds prime importance and is above all the elements. The community and support that Omdena provides are fascinating. Throughout the entire project, we were guided by an experienced mentor that not only supported us but provided us with a proper roadmap whenever we were baffled.
Thanks to collaborators for the opinions, diversity and support brought to the table during the project which made me think of taking an enormous step of becoming task manager!
I was an active collaborator in other tasks before I came up with the idea that there were too many features during the initial stage so we need to extract some useful features to train the model better. This is where I started with Task: Feature Extraction. I had little knowledge of this earlier so I thought really twice before taking this tremendous step. But in the end, it is not necessary that the task manager should always know all the answers because it is always about learning and helping each other grow.
Throughout the project and in Task 3 specifically, I learned AWS Cloud Services, Flask, Spacy, Sentiment Analysis, Risk Score Prediction, NLTK, Lemmatization, Topic Modeling, Knowledge Graph, Lasso Regression, BERT, BART, XLNet, and GPT2. All these high-level words and coding were not even my forte but by working on the tasks I learned a lot of novel concepts. As a task manager of the task, I was supposed to hold weekly meetings with fellow collaborators to have discussions on how to move forward with the task and present the progress done to the mentor every Saturday.
Though I was less experienced than some team members, I was always learning from them. I am very grateful to have supportive members and collaborators while working on the project that believed in me and helped me to enhance my Data Science skills.
I believe that efforts never go in vain because the overall final model that we prepared consisted of lots of ups and downs. Some things we modeled were not really included in our final deliverables, but we learned to implement that by reading research and blogs published on it. Another massive takeaway from this project is that we can easily overcome data problems in the initial stage by working with an ardor to scrape more data and continuously ponder over the Internet. In this project, we were not having much data to train the model, so we all started from the very first step by collecting data and then finally deployed it online for the parent organization.
After this project experience, I can’t go back to the old normal because Omdena showed me the power of collaboration and innovation that can prove beneficial in real-life too. I am glad that I also got the opportunity to join this other upcoming project-Improving the Lives of Cancer Patients by Identifying Existing Non-Cancer Generic Drugs.