Ahmedrufai Otuoze

Led design and implementation of Azure-based data architecture for NLNG Supply Chain Digitalization, building real-time ingestion and processing pipelines and reducing release cycles from weeks to days via CI/CD automation.

Curated and annotated large Python and JSON datasets for LLM training, developed annotation schemas and automated analysis tooling, and applied PEFT techniques (LoRA/QLoRA) to fine-tune models for code tasks.

Developed prompt templates, integrated new language models into the Unify API, and conducted model evaluations and benchmarking to improve LLM Router performance and user guidance.

Supported web platform analytics and reporting across projects in Uganda, Mozambique, and Ghana using PowerBI and SEO improvements to optimize user experience and campaign performance.

Delivered cloud ETL, real-time analytics, and ML-driven predictive maintenance for NLNG, migrating warehouses to Azure and reducing unplanned downtime by 35% while improving reporting performance and operational efficiency by ~30%.

Mentored 13 students in Python fundamentals, data wrangling with Pandas and Statmodels, and introductory machine learning with Scikit-learn to build foundational data analytics skills.

Optimized ETL with Airflow and Spark to process multi-terabyte financial datasets, built Snowflake/Redshift models and real-time analytics using Kafka/Druid, reducing latency 30% and lowering cloud costs by 20%.

Led design and deployment of Snowflake-based ELT pipelines and analytics dashboards, implemented Airflow automation and DBT transformations, reducing pipeline latency 30% and improving customer retention by 15%.

Built scalable data ingestion and processing systems using PySpark, Kafka, AWS (EC2, Lambda, S3), Snowflake/Redshift, and containerized services, processing large-scale IoT and transactional datasets and improving data pipeline reliability.

Bachelor-level studies in Design and Production Engineering completed in 2019.