Seed Week and Vibeathon at Bversity became much more than just another academic event, it turned into a real test of teamwork, clarity and problem solving. What started as a confusing challenge slowly transformed into a project that pushed our three learners to think, build and grow together. From collecting oncology datasets to shaping a knowledge graph and finally bringing an AI assistant to life, every step taught something new. This blog takes you through the OncoGraph journey, told in three voices across different phases. It is a story of how curiosity, patience and a few well timed insights can turn chaos into something meaningful.

"Every error we faced is simply experience, shaping us for success."

I can still remember the moment everything finally clicked while building Oncograph. All the confusion we had at the start slowly disappeared after Philip Moses explained LLMs, nodes, edges and knowledge graphs in a way that suddenly made everything feel possible, and that was when Eric, Ammathullah and I felt ready to take it on. The whole journey took just two days, but it felt like an adventure. On the prep day before Vibeathon, we were honestly ready to give up because the concepts felt overwhelming, yet Rahul’s thought process tips and Philip’s steady guidance pushed us to show up the next day with the simple intention of learning. When the Bversity team started the timer, it gave us an instant spark. As we collected cancer datasets and built the knowledge graph together, completing that first milestone filled us with confidence and joy.

- Brindha T M

"Even the hardest challenges turn surprisingly fun when you lean into them the right way."

Our team, Eric Johnson, Brinda, and I, was tasked with building Oncograph, a biomarker knowledge-navigation agent designed to solve the challenge of fragmented oncology data by integrating a knowledge graph with an AI reasoning layer. We began by collecting key datasets from NCBI, OncoKB, PDB, and Ensembl, compiling everything into CSVs. The next major step was merging this information and shaping it into a coherent knowledge graph in JSON.

Then came the fun and chaotic part, building the interface with GitHub Copilot. Since we were new to it, our prompts had to be painfully precise. After multiple attempts, endless debugging, and  patience, we managed to build a basic functional interface and  linked it to our knowledge graph.

- Amatullah Rampurwala

"Clarity cuts through chaos, turning every moment into a lesson worth keeping."

The real adventure began when we moved into app-building: wrestling with GitHub Copilot, debugging what felt like a hydra of errors, and slowly learning how to “talk” to the LLM so it actually behaved. Piece by piece, we got our agent to answer multi-hop oncology questions and built a live graph visualization to show its reasoning. Presenting it to the judges, especially Agilisium’s Founder & CEO, Raj Babu, was nerve-wracking, but when asked what a domain scientist brings to a tech-heavy build, I could confidently say the graph itself depends on deep biological understanding. Seeing the judges nod, and watching our fully working app win the cash prize at the end, made the chaos, stress, and fun of those eight hours feel absolutely worth it.

- Eric Johnson P