July – Sept 2026 · FlyRank AI
FlyRank Capstone: Content Decay Prediction and Search Intelligence
Automated model that evaluates content decay risk and guides proactive asset refreshing. Built and evaluated ML ranking models and data pipelines across ~79M rows of search discoverability data.
Problem
Search visibility decays over time as query intent and competing pages shift, making it difficult to identify which content assets are at risk before traffic drops.
What I Built
Developed an automated machine learning ranking and content decay prediction pipeline during my ML Engineering Internship at FlyRank AI, processing ~79M rows of search discoverability data using Python, DuckDB, and Hugging Face.
Result
Evaluated content decay risk to guide proactive asset refreshing and presented the research at Prolog EXPO 2026, earning 2nd place.