I have done Bachelor’s in Statistics and a master’s in computer applications. An experienced IT professional with over 22 years of playing multiple technical and managerial roles, always looking for challenges and constant learning and upskilling, which helps the individual and organization mutually. Before joining the PGP-AIML Course, I was managing the support organization for Zuora India Pvt. Ltd. They are the leaders in subscription-based billing.
Data services and analytics is my passion, and with a statistics background and upskilling based on industry trends, the idea of learning AI/ML was the need of the hour.
The key objective of the Capstone Project is to apply our learning and work on a use case that is relevant to the current context and covers multiple areas of our learning. We chose Multi lesion detection of X-ray images which solves the industry-wide problem for radiologists. We applied ML and AI techniques for this image classification and detection problem and learned multiple new techniques in the process. The results were helpful and comparable with the best in the industry as well. The results were benchmarked with the best researchers who worked on this problem earlier.
The key pain points I faced during the project were the availability of computation-intensive heavy systems, and we had to hit roadblocks multiple times. We then strategized to solve the problem in batches which took a lot of time and energy. We met every week and worked with the mentor. Some of the experiments and methods we tried had almost no reference on the internet as well; we had to dig deep and solve them ourselves.
The mentor played a key role and provided references from his personal experiences and asked us to try relevant stuff; when we are trying as a team, it is critical to have a guide who is sharp, focused, and experienced. He also helped us scope out the problem so we could complete things on time. And with the Capstone project, we benchmarked with the industry’s best researcher, and for a few lesions/diseases that we detected from the X-ray images, the accuracy was much higher than the benchmarked values. This solution and research are pretty useful to radiologists, and further can be extended.
The capstone project helps the learner to apply all the learning and explore options to solve real-world problems confidently. Spending quality time with the team and mentor has helped. We were able to learn multiple up sampling, down sampling, and additional new metrics which were never used before for results and evaluation. The learning experience was tough but enjoyable. Personally, it has helped me improve patience as each model takes a sufficient amount of time to run and execute. At the same time, it helps work on multiple things and monitor each of the activities well to ensure correctness. It was interesting that I learned how to execute machine learning models. Professionally it helped improve the visibility with corporates, and I landed an engineering role for a data services team. Soon we are planning to work on AI and ML models as well.
The advice to the new learners would be to be patient and learn well. Learning should be an enjoyable experience. Results will eventually happen. Thinking of the job/monetary benefits in mind will not help in learning. Focus on peer learning and work on improving the network and collaboration.
