At Akumen by Qubiqon, I’m training in data engineering workflows across Python, SQL, data modelling, and pipeline orchestration. I’m also working with Azure Data Factory, Azure Databricks, Microsoft Fabric, Apache Spark, and Power BI.
I built WILDLIGHT, an environmental audio classification pipeline using ESC-50 and a ResNet18 model. I implemented shared preprocessing for training and inference, ran SpecAugment experiments and five-fold cross-validation, and developed a Gradio interface for testing predictions.
I also built a multi-stage AI research workflow in n8n that collects AI news, extracts original articles, gathers search-based evidence, and produces structured outputs. I explored local LLM integration and topic-based routing.
Alongside my data work, I’ve spent six years producing music and designing sound across dark R&B, Afro-influenced music, and experimental soundscapes. I’m exploring Genesis, an AI-assisted audio concept connecting sound design, audio technology, and interactive experiences.

