Internship Description

Thu Tra Pham,'23SPS, was a data analyst intern at Endear. Endear is developing an AI-based voice-recognition chatbot designed explicitly for people who need mental health support. Thu trained and supervised learning models and helped Endear prove the concept of automating a few humanmachine conversational interactions following standard practices in the mental health industry.

This summer, I gained valuable experience working as a data analyst intern at Endear. During my internship, I had the chance of working on a variety of projects and tasks, utilizing my data science skills. Endear is a startup focused on leveraging artificial intelligence and machine learning to connect individuals dealing with mental health issues to the available support resources. My role involved training supervised learning models and contributing to the development of an AI-powered chatbot tailored specifically for individuals seeking mental health assistance. My main task was to engineer sentence classifiers (fastText) to predict conversation-end for accurate response rate metrics. Within our team, the goal was to demonstrate the feasibility of automating certain human-machine conversations using standard practices in the mental health sector. My input has helped Endear explore how technology can be utilized to initiate conversations and assess the initial treatment requirements of underserved communities, addressing the care gap caused by a shortage of licensed human caregivers for those in need.

During my fellowship with Endear, I had the chance to delve into complex health analytics problems. The internship provides an invaluable opportunity to utilize the data analytics skills I have acquired at Columbia University into real-world applications. Columbia courses, including Research Design, Applied Analytics Frameworks, and Machine Learning, have helped sharpen my technical skills drastically. I have built automated reports in SQL and PowerBI, transforming raw data into actionable insights. I have also used machine learning to develop sentence classifiers to enhance the accuracy of response rate measurements, as well as some research in human behavior. Moreover, the skillsets I have applied to this fellowship go beyond technical skills. I also orchestrated and led cross-functional meetings, fostering collaboration among the engineering and product team. My learning at Columbia has not only deepened my expertise in data analytics but also expanded my understanding of the transformative capabilities of AI-driven healthcare solutions.

During my time at Endear, I faced a few challenges. Since Endear is an early stage startup, the internship required me to wear many hats. While my primary role was as a data analyst, I was also involved in tasks ranging from wireframing to feature development. Balancing these diverse responsibilities while maintaining a high level of quality was a significant challenge. The working environment at Endear was fast paced. I had to quickly adapt to changing priorities and sometimes pivot in response to new insights. Flexibility was essential. In the context of creating an AI-based product, there was limited data available for current events. This meant that I had to employ creative approaches to data collection and model training. Despite these challenges, they provided valuable learning opportunities and contributed to my growth as an aspiring data analyst. I am grateful for the opportunity to contribute to empowering the future of digital health care through my internship at Endear.

Through my internship at Endear, I gained a deep understanding of the evolving health care landscape, particularly how the health care system operates. It has been eye-opening to learn about the pressing issues, such as the severe shortage of mental health providers, which unfortunately results in 30 million Americans going without treatment. The fellowship has taught me the challenges of operating a digital health startup, and how AI and machine learning can play a pivotal role in closing the care gap created by the shortage of licensed human caregivers for people in need. My lifelong aspiration is to make a meaningful difference in the lives of disadvantaged individuals. I am committed to continuing to apply my data science skills toward the development of a communication tool tailored to those in need of mental health support. Moreover, I aim to facilitate connections between potential patients and licensed caregivers, enabling them to access essential consultations and treatments.