IVS Hackathon 2026: Solving quality in the AI era
AI can accomplish remarkable tasks, but who will guarantee the quality of AI? IVS Hackathon 2026 has officially launched as an “open-ended problem,” bringing IVS members together to seek answers to the core challenge of AI Testing: measuring and validating the quality of AI Agents in real-world applications.
On May 25, IVS officially kicked off the Hackathon competition, drawing the participation of 16 teams from three regions. More than just a tech playground, this year’s Hackathon is designed as a journey of discovery, where individuals can push their limits while contributing practical solutions to the field of AI Testing.
Speaking at the kick-off event, Ms. Phan Thi Vi Ba – FSUL of IVS, emphasized her expectations: “The Hackathon must go beyond mere concepts. It needs to generate actionable innovations that will shape IVS’s AI Testing capabilities in the upcoming period.”
The competition consists of three rounds with a transparent screening and development roadmap. Following the first two rounds, the finest teams will gather in Hanoi in early July for the Grand Finale. Here, teams will not only finalize their demos and presentations but also prove their practical value by deploying their use cases into at least two real-world accounts within the unit.
The focal point of this year’s Hackathon revolves around a critical question: as AI becomes increasingly automated and intelligent, how will the quality assurance of AI Agents themselves be executed?
Driven by this question, IVS has set a clear objective: to build a framework or an agent capable of testing and evaluating the quality of other AI Agents. This is not only a technological challenge but also a methodological puzzle, as testing transitions from traditional systems to adapt to AI-native models.
Based on this strategic direction, specific tasks assigned to the competing teams include:
- Designing a testing methodology for AI and Agents;
- Developing a playbook for validating AI-native systems;
- Defining a harness-based quality evaluation framework;
- Developing a Validator Agent – an “AI built to test AI”.
The differentiator of this Hackathon lies not only in its problem statement but also in its approach: shifting from a traditional testing mindset to a systematic and scalable AI validation mindset. This marks a vital step for IVS to progressively build battle-tested AI Testing capabilities, staying ready to meet the ever-rising demands of the market.
Therefore, IVS Hackathon 2026 is far more than just an internal competition; it is the birthplace of pioneering methodologies. As ideas materialize into solutions, and solutions evolve into battle-tested capabilities, IVS will lay a solid brick in FPT Software’s overarching AI Transformation strategy.
