Henry-G Tanoh

Architected large-scale ETL pipelines using Hadoop, Spark, and SQL to process 10TB+ daily attribution data, ensuring data accuracy across distributed systems for Europe's leading ad-tech platform. Implemented a low-latency real-time bidding system with Apache Flink for automated bid resets, improving campaign budget utilization by 5% through intelligent optimization. Designed and implemented an A/B testing framework for streaming applications using Apache Flink, HDFS, and Kafka, enabling real-time experimentation and validation of new features across high-throughput data pipelines. Built monitoring infrastructure for ML forecasting APIs using MAPE/SMAPE metrics and Tableau, tracking prediction accuracy for models serving millions of daily requests.

Engineered edge-to-cloud data pipeline in Go/C++ processing terabytes of IoT sensor data from 100+ robots, with scalable AWS S3 storage architecture for real-time analytics. Built a CLI toolkit and automated deployment system, reducing installation time by 70% and data retrieval from minutes to seconds. Technologies: C++, Go, Python, AWS, Linux, Windows, Docker, Apache Spark, Pyspark, AWS SWF

Architected a high-throughput ingestion system using HDFS for distributed storage, processing terabytes of daily vehicle sensor data. Implemented partitioning and compression strategies that reduced storage costs by 35%. Built real-time streaming pipelines with Apache Spark for multi-sensor automotive data, ensuring sub-second latency for critical safety telemetry. Designed data pipeline with AWS EMR and IoT Core, capturing real-time telemetry from test vehicle fleet, enabling rapid analytics for engineering teams.