Engineering
Intelligence
Across Every
Layer.

Trained in industrial control, automation and sensor-driven systems — the engineering discipline that underpins everything I do with data today.

Delivered machine learning solutions for health informatics within a cross-functional AI team. Led an end-to-end ML pipeline that cut training and deployment compute costs by roughly 90% across two core models while lifting recall at the target precision threshold by up to 10%. Designed feature-engineering algorithms that cut memory use by up to 32× for high-cardinality categorical features in tree-based models.

Designed and built a smart energy switching system integrating IoT sensors, control logic and power electronics.

Managed end-to-end electrical installations, diagnostics, testing and commissioning across residential and commercial projects, ensuring compliance with Nigerian safety standards and building codes. Produced inspection and progress reports documenting installation status and defects for project stakeholders and senior engineers. Interpreted technical drawings and specifications to coordinate on-site works across multiple concurrent projects.

Annotated and reviewed datasets supporting Large Language Model, NLP, Computer Vision and Generative AI development. Ran human-in-the-loop evaluations, scoring AI-generated output against project-specific quality criteria and annotation guidelines. Ranked competing AI responses for relevance, accuracy, helpfulness and clarity as part of RLHF response comparison. Performed Named Entity Recognition, sentiment analysis and text classification, plus dataset validation and quality audits.

Wrote and published how-to guides, tutorials and software documentation, translating complex product features for technical and non-technical readers. Edited documentation for clarity, consistency and usability, aligning it with product functionality and brand standards.

Evaluated Automatic Speech Recognition (ASR) output for transcription accuracy, completeness, grammar and contextual correctness. Performed speaker identification and timestamp synchronisation to raise speech dataset quality. Flagged transcription errors, inconsistencies and annotation defects for correction against detailed project guidelines.

Built AI agent systems on the Claude and OpenAI APIs for business workflow automation. Automated multi-step business processes in n8n, including email routing pipelines, agentic task orchestration and multi-trigger event systems.

Delivered structured written product feedback for global brands across fintech, healthcare and e-commerce, in research supporting large language model fine-tuning.

Annotated 1,000+ short video clips for AI training data, describing actions and events in egocentric (first-person) footage. Verified AI-generated annotations against source footage on the Human Verifier track, correcting every label that failed project accuracy criteria. Produced segment-by-segment annotations from scratch, holding consistent labelling quality across multiple concurrent task types and distinguishing subtle differences in movement, object interaction and action sequences.

Authored and refined prompts for YouTube's AI search mode, improving the relevance of returned results against user query intent. Assessed model output against prompt intent and reworded prompts wherever responses missed the mark.