Dimitris C.

Sole Data Scientist in a fast-paced IoT startup, developing predictive models for machine diagnostics and optimization. Built classification models achieving 95%+ accuracy for traditional coffee machines using water flow and extraction data. Actively contributed to product development by analyzing IoT data and proposing actionable insights, new features, and algorithmic improvements. Created sensor data models, improved device performance and enhanced accuracy. Designed and implemented internal tools including:

• • Web Analytics App: Real-time monitoring tool for coffee machine performance, integrating firmware simulation

tracking, ML-based status classification, and algorithms to assess device functionality and anomalies.

• • GPT-powered Chatbot: Natural language interface for

accessing analytics.

• • Coffee Classification System: Methodology based on flow for predictive maintenance (internally documented). Collaborated directly with Product and Engineering in Agile sprints to deliver ML-driven features. Delivered company-wide reports, enabling improved decision-making across departments. Designed custom B2B dashboards (e.g. for service maintenance, quality monitoring), resulting in 30% improvement in coffee quality. Designed and led the integration of next-day reporting within the customer platform, enabling B2B clients to access operational and performance metrics Responded to ad hoc requests and engaged with B2B customers to understand operational challenges and inform the development of tailored analytics solutions. Designed and owned the logic behind classification algorithms for device status and diagnostics before feature deployment. Full ownership of the Data domain, including internal reporting architecture, internal analytics tools , and ML pipelines.

Portfolio analysis and financial modeling to support securitization strategies and drive stock reduction of Non Performing Exposures (NPEs). Implemented Allocated Collateral Real Estate Valuation logic, enhancing loan assessment accuracy and recoverability projections. Developed a Cashflow Algorithm Model to simulate NPE loan amortization based on user-defined rules and restructuring terms. Defined data requirements and validated outputs while coordinating external analytics teams toward banking objectives.

Proposed new BI features and data-driven improvements based on proactive analysis and collaboration with Product teams. Delivered BI dashboards for IoT, energy, and smart bin sectors using Power BI. Implemented supervised and unsupervised ML models to optimize onboarding and cluster customer behavior. Designed backend solutions for B2B clients including NoSQL databases, enabling integration of third-party device data. Assessed Apache Spark for data processing and platform scalability.