Data analytics internship
In July, we welcomed Zahra Safavifar to the Data Analytics DOCOsoft team. Zahra is a Researcher Ph.D. Candidate at ML-LABS- SFI Centre Research Training in Machine Learning. We met with the Zahra to take a brief look into her experience as an intern in DOCOsoft over the last 5 months.
Can you tell us a little bit about yourself?
My name is Zahra Safvifar. I am Iranian and study computer science at University College Dublin. I am in the third year of my Ph.D. program. As part of my ML-LABs program, I have done an internship at DOCOsoft. I have 7 years of work experience in BI and big data engineering in Iran. DOCOsoft was my first industry experience in Ireland, and I would say I had a very positive experience working with a friendly and supportive data analytics team.
What projects did you primarily spend your time working on?
I worked on a claim processing optimisation project. I worked on using machine learning for task allocation in the claim handling process to improve handling process speed.
What has been your biggest achievement throughout your internship?
I implemented a state-of-the-art Reinforcement Learning algorithm for task allocation in the claim-handling process. The primary results show that the current algorithm can save 54 working days in comparison to the current task assignment method.
What was your favourite thing about working in the Data Analytics team?
I was independent, and the lead and other team members were very supportive. Also, I liked DOCOsoft social events.
What element of Data Analytics did you find most interesting?
Discovering insight from a lake of messy numbers and text is a very interesting task. In addition, using machine learning for solving a real-world problem is a fascinating task.
What would be your advice to any future interns starting at DOCOsoft?
DOCOsoft can be a good place to start your career since the team is not too big and you can feel you are part of a family.
If you are interested in an internship or career at DOCOsoft, visit our careers page.