Abstract
My technical project and STS research paper are connected through the broader question of how complex technical systems depend on the relationships between human users, organizational goals, and non-human technical actors. My STS research examines the Ariane 5 Flight 501 failure through Actor-Network Theory, and how human leadership acted as the network builder to influence technical decisions and actors to neglect proper safety protocols. My technical work focuses on designing a chatbot-based interface that helps nontechnical users interact with a vertical microfarm system – as the network builder I assembled a network of both human actors, like Babylon Microfarm engineers, and nonhuman actors like software packages and the hardware inside the farm. Although these projects differ topically, both show that technology cannot be understood as isolated code or hardware. Instead, technical systems succeed or fail based on how well their networks of users, designers, assumptions, and machines are aligned.
My technical project proposes a text-based conversational interface for Babylon Micro-Farms. The goal of the system is to let users such as students, chefs, and office workers ask questions about a microfarm’s status and issue commands without needing deep technical knowledge. The project improves on the existing mobile app, which offered limited support options and required users to rely heavily on company representatives to troubleshoot. By using an LLM-based chatbot, the system could translate natural language into farm-related actions or status queries. This design required software engineering decisions about modularity, databases, user personalization, privacy, and secure communication. In this sense, the project was not just about building a chatbot. It was about making a complex agricultural technology more accessible to the people expected to use it.
My STS research paper examines the 1996 Ariane 5 Flight 501 launch failure. I argue that the failure was not simply caused by a software overflow, but by a broader sociotechnical network built around software reuse and insufficient validation. Using Actor-Network Theory, I identify Ariane 5 program leadership as the network builder that commandeered engineers and made the decision to reuse the Ariane 4 software and create a tightly coupled launch architecture, as a result of organizational priorities. This framework shows that the crash resulted from the interaction of technical components and institutional decisions, not from a single defective line of code.
Working on both projects together helped me see technical design more critically. My technical project pushed me to think about how software can make complicated systems easier and safer for users. My STS research reminded me that design choices also carry hidden motives and assumptions, often by human actors, and this can also lead to developers trusting existing systems as reliable black boxes. Together, the projects showed me that responsible computer science work requires more than functional code. It also requires understanding the social context and organizational pressures, as well as the technical dependencies that shape how a system behaves in the real world.
