At Espatika, I translate business questions into recommendations for non-technical stakeholders, completing analytics projects for SME clients. I’ve built classification and forecasting models across 100K+ records and automated workflows that reduced analysis and reporting time by 30%+.
I’ve also developed and deployed LLM-powered chatbot agents, taking AI solutions from prototype to production by integrating LLM APIs with Python and website technologies.
For a Data Science Industry Project with Deep Planet at Queen Mary University of London, I led a four-person team developing wind-power forecasting models from turbine SCADA data. Our deep-learning models achieved up to 11.0% skill at a 60-minute horizon.
In my projects, I built an end-to-end loan-default MLOps pipeline and analyzed Apple-related tweets around the iPhone 3GS launch. The work included model fairness and drift audits, and identified a 43-percentage-point increase in negativity toward the iPhone OS 3.0 update topic.
