Cloud & Solutions Architect

Owned architecture and technical decision-making for approximately 15 PRD-driven product initiatives within a 14-person engineering team. Designed a multi-cloud data platform consolidating Firebase and AWS S3 device, user, and event data into BigQuery, serving 300,000 users and 250,000 active connected appliances across more than 100 tables and 180 TB of data.

Designed a data lifecycle strategy that reduced BigQuery costs by 50% and optimized Cloud Run configuration to reduce costs by 75%. Provisioned test and production infrastructure with Terraform and Git-based workflows. Led Vertex AI and Gemini adoption through an internal chatbot, AI coding assistant, automated reviewer, and LLM report summarization; delivered usage dashboards and a device-based audience segmentation model for 5 million Smart TVs; mentored one intern and four junior engineers.

Designed and developed the serverless backend for Vestel's Home Connectivity Alliance integration, enabling bidirectional cross-brand control of connected home appliances. Built Node.js discovery, status, and command services using AWS Lambda, API Gateway, DynamoDB, and S3.

Delivered integrations with Samsung, LG, and Arçelik devices while meeting an approximately five-second end-to-end latency requirement for command and status flows. Provisioned infrastructure with AWS CloudFormation, monitored service logs through CloudWatch, used PostgreSQL where required, and wrote unit tests with Jest.

Set up the organization's data pipeline and refactored SQL transformations to improve the reliability of internal reporting.

Built and operated a monthly sales forecasting model for 120 SKUs, achieving approximately 70–80% forecast accuracy using SQL and SPSS.

Developed and maintained two AWS-hosted energy trading and smart energy management platforms using Django: EnergyPool.live for power plant consumption and market data, and Epiq.live for plant production, consumption, and revenue reporting.

Maintained Python background services for electricity market price and power plant data ingestion, resolving production issues and ensuring service stability. Built end-to-end ML forecasting workflows with data extraction and QuickSight dashboards, achieving about 92% accuracy for electricity price forecasting and 80–90% accuracy for plant production and consumption forecasting.

Developed a Python-based network discovery tool for identifying devices connected to a network.

Implemented SSH connectivity to retrieve and report device hostnames and IP addresses across the discovered network.

Bachelor of Science in Electrical and Electronics Engineering.

Erasmus exchange program completed during Spring 2017.