Daniel Yeke

I build production RAG architectures, LLM orchestration systems, and edge ML pipelines. My work includes a multi-provider LLM router and self-correcting RAG system designed to reduce API-limit risks, control cost, and avoid single points of failure.

I built automated RAG evaluation and tracing workflows with Ragas and LangSmith, monitoring faithfulness, relevance, context precision, latency, token usage, and generation drift. I use these signals to improve prompts, chunking, retrieval, and model behavior.

I also designed an autonomous sensor-fusion rover for landmine and anomaly detection, optimizing a Random Forest model for real-time Raspberry Pi inference through feature selection and ONNX quantization.

As Lead Technical & Asset Systems Consultant at AuditFlow System Ltd / BIPC Nigeria, I spearhead technical system audits, data-integrity architectures, asset-evaluation workflows, scalable infrastructure guidance, and automated verification pipelines.