Sonoi Ogbuefi

Built an auditable, LLM-assisted crypto trading and risk platform using LangGraph, RAG, Redis, and unified REST/WebSocket integrations across multiple exchanges. Implemented deterministic pre-trade controls, resilient asynchronous execution, and controlled testing workflows to support safe institutional trading.

Built scalable backend and workflow infrastructure for developing, evaluating, and deploying LLM applications such as customer-support assistants, document Q&A systems, and extraction workflows. Implemented RAG, prompt orchestration, evaluation, versioned deployment, and production monitoring using Python, FastAPI, RabbitMQ, Redis, and MySQL.

Built production-grade anomaly-detection and risk-prediction systems using Python, TensorFlow/Keras, scikit-learn, and XGBoost, delivering reproducible training and batch/real-time inference with FastAPI. Productionized and monitored models using Domino, Docker, Kubernetes, Git, and CI/CD, supporting scalable, reliable deployment and drift detection.

Built scalable feature-engineering, model-training, and batch-inference pipelines using PySpark, Scala, Spark SQL, and MLlib. Optimized distributed workloads and productionized models through Databricks notebooks, scheduled jobs, validation checks, and retraining workflows.

Developed Caffe-based face detection and image classification models using Python, then integrated and optimized them for real-time Snapdragon camera pipelines with C++, OpenCV, ARM NEON, and Hexagon DSP/HVX.

Earned an M.S. in Computer Science from Florida Institute of Technology between 2019 and 2020.

Grade: -

Activities and societies: -

Earned a B.S. in Computer Science from Caritas University between 2012 and 2016.