PromptWars x Parul University: Inside a Google-Backed Vibe-Coding Hackathon

At 9 a.m. on 5 September 2026, the Bhagat Singh Block at Parul University filled with laptops and focused faces. The event was PromptWars, and its brief was unlike a normal hackathon: students would receive a surprise problem and just a few hours to build a working AI solution using nothing but prompts. This was vibe coding, and it drew a crowd.

PromptWars wasn’t any small affair. It was designed for engineering and BCA students; it was run in partnership with Google for Developers, Hack2Skill, and the Build with AI initiative, giving students a taste of industry experience. The single, clear brief: turn plain-language prompts into a functional AI application, under time pressure, on a problem nobody could prepare for in advance.

Perhaps the most telling detail was the turnout. 510 students registered, and more than 300 showed up in person, entirely through organic interest. No paid advertisements ran; no external agency handled outreach. Just one simple move – it was spread through student networks, class groups, and conversations. That level of voluntary, genuine participation says something real about the appetite for hands-on AI challenges within Parul University’s CSE and AIML community.

The event was a mix of effective student leadership and faculty support. Aman Sayyad, a second-year student of CSE (AIML) at PIET, was the lead organizer who managed the proposal, coordinated with the faculty and external partners, managed the volunteers, and was on the ground from the initial meeting to the final assembly of the leaderboard. The institutional guidance was provided by Dr. Kamal Sutaria, Head of CSE (AIML), and Assistant Prof. Ms. Ritu Agarwal was the academic support provided. The assistant professors Ms. Nidhi Shah and Mr. Siddhart Jain were the Judges for the event.

At the start of the event, the problem was announced, and it was a clever choice of problem. The participants were required to develop an AI agent or website that would serve as a mentor to final-year students when selecting and designing their final-year projects. The system was required to recommend proper project topics amongst the available projects in the department, suggested technologies, difficulty level of the projects, and the available time for the students; and also to create a complete plan, architecture, dataset suggestions, and week-wise time frame for the project, which students could save offline. The short was a reflection of an actual problem: many final-year students spend weeks looking for a project that is suitable for their skill set and deadlines. The rules were also strict: files with less than 10 megabytes in size, only 2 or 3 tries to submit (the final being the important one), and quality, security, and accuracy were scored.

Across the hall, students worked the way this new format demands, not by writing code line by line, but by crafting prompts and iterating on the output. Tools like Google Antigravity, Gemini, DeepSeek, and others appeared across many projects. Several built complete front ends and back ends in roughly two hours. The great equalizer was clear: prior coding background helped but did not guarantee a high rank. What decided placement was the ability to understand the problem, frame a sharp prompt, and test the result, so first-year students competed on level terms with seniors, and the leaderboard mixed years and experience levels freely. As several participants noted, the quality of the output depended more on prompt craft than on hours of coding.

The live leaderboard showed how tight the contest was, with the leading scores separated by fractions of a point, with Dikshant Agarwal topping the numerical ranking. The top ten were then invited to pitch their working websites to the jury, and the projects impressed. Dikshant Agarwal’s IT Final Year Project Suggestion and Data Management System offered role-based logins and exported roadmaps, even generating likely viva questions. The three winners, decided on both scores and live pitches, were

• Winner – Dev Sonar, “Project OrbitMentor,” a dual-engine AI architect that turned a student’s skills and interests into scoped project blueprints and a 16-week roadmap, built with a React and TypeScript front end and a Python back end calling the Gemini API, planned with a full requirements document before any code.
• First runner-up – Aryan Maurya, “ProjectPilot,” focused on rapid idea-to-product conversion in an AI-first setting.
• Second runner-up – Harsh Chavan, “Meridian,” a functional AI prototype built and deployed with Google Antigravity in a single day.

Several participant projects went live for anyone to try, including one AI project-mentor website and the event’s own build.

The future is being built by those who know how to code it. Take your place in the tech world with B.Tech CS AI & ML, B.Tech in Information Technology (IT), Bachelor of Computer Applications (BCA), and M.Tech in Computer Science & IT at Parul University.

PromptWars was more than a competition; it was a snapshot of where software is heading and of a student community ready for it. It showed that large-scale student activity can run on organic interest alone, that a Google-backed, industry-relevant format can be delivered by student leadership with faculty support, and that in an AI-first world, clarity of thought and prompt skill can matter as much as years of coding. For Parul University’s CSE and AIML students, it was a hands-on introduction to the tools and mindset that are reshaping the profession they are about to enter.