Ticketmaster
Ticketmaster, a global ticket sales and distribution leader, is renowned for its robust platform that handles millions of transactions daily. In early 2024, I joined Ticketmaster for an eight-week rapid engagement through Artium to develop a prototype application to revolutionize certain aspects of their operations. Utilizing cutting-edge technologies and methodologies from my venture, Dev3loper.ai, we quickly began delivering value from the first week.
A Continuous Integration and Continuous Delivery (CI/CD) pipeline was set up immediately. This enabled us to develop a working Node.js application with a React front end, which we quickly iterated on throughout the engagement. Leveraging expertise in Generative AI, software development, and extreme programming, we integrated advanced AI both in the development process and within the application itself. This engagement underlined our proficiency in Large Enterprise, Financial Technology, Software Development, and Generative AI.
Rapid Engagement: Event Discovery
Project Overview
Engagement Period: January 2024 – February 2024
During this eight-week rapid engagement at Ticketmaster, I led the development of a prototype application focused on Event Discovery. Although the product details remain under NDA, the project's overarching goal was to leverage AI and modern development practices to create an impactful solution.
Technologies and Applications
The technological stack utilized for this project included:
- Frontend: React (TypeScript)
- Backend: Node.js with Express (TypeScript)
- AI Integration: OpenAI API for Chat Completions, facilitating a Retrieval-Augmented Generation (RAG) architecture
- Database: Pinecone, a Vector Store Database as a Service
- AI Tools: GitHub Copilot integrated into the IDE for enhanced development efficiency and as a pair programming partner
Team Dynamics
Our team comprises four engineers from Artium who work directly with stakeholders at Ticketmaster to refine requirements in real-time. This collaboration ensured the prototype aligned with the stakeholders' vision and business needs.
Impact and Practices
Given the short engagement period, we employed highly iterative development cycles. We began with weekly iterations but quickly transitioned to daily cycles in the last few weeks to rapidly identify and implement the most valuable features for the customer.
A core aspect of our success was the implementation of Test-Driven Development (TDD) and Extreme Programming (XP) principles, foundational to my approach at Dev3loper.ai. These practices ensured robust, reliable code from the outset, with an iterative development process allowing continuous improvement.
I introduced the idea of "Repeat Tests" to help mitigate non-deterministic outcomes of integrating LLM technology into a production application. This process involves running a suite of tests multiple times, such as 10+, and aiming for a success rate above a certain threshold, like 80%. We discovered that minor tweaks to embedding the data or changes in the prompt could dramatically affect the product's quality. Using this process, we ensured that any changes to the application, in conjunction with our traditional extreme programming approach, met customer expectations and produced outcomes that delighted users.
Our innovative use of GitHub Copilot significantly boosted personal efficiency and streamlined the coding process. Acting as a virtual pair programming partner, Copilot provided real-time suggestions and facilitated seamless coding workflows.
Additionally, integrating the OpenAI API for Chat Completions enabled the creation of a sophisticated RAG architecture. We conducted numerous experiments, iterating through various prompt changes to fine-tune the AI's performance, leading to significant application functionality enhancements.
Outcomes
The prototype application received a very positive reception during user acceptance testing. The internal team at Ticketmaster approved the prototype for further development, and plans are underway to engage with us soon for the next stage of development. This positive outcome underscores the impact and viability of the innovative solutions developed.
These contributions align with the services offered by Dev3loper.ai, showcasing our expertise in AI-driven solutions, accelerated prototyping, and agile methodologies. The project's successful transition from proof of concept to a potential full-scale development illustrates the effectiveness of our approach.
Conclusion
Working with Ticketmaster was an exhilarating experience highlighting the power of rapid prototyping and AI integration. By applying extreme programming principles, TDD, and advanced AI tools, we delivered a prototype that met and exceeded expectations. This engagement serves as a testament to the capabilities of Dev3loper.ai and our commitment to delivering high-impact, innovative software solutions.
Get In Touch
We'd love to hear from you! Whether you have a question about our services, need a consultation, or just want to connect, our team is here to help. Reach out to us through the form, or contact us directly via social media.
