Building production LLM systems and agentic AI architectures that turn complex data into real-time, actionable insight. Grounded in a research background in distributed and federated learning, now focused on shipping ML and GenAI systems that hold up in production.
I'm a Data Scientist and AI/ML Engineer with 3+ years of experience developing and deploying machine learning and LLM-based systems across healthcare, government, and industrial IoT sectors.
My expertise spans from research-based distributed AI systems to production-ready agentic architectures and MLOps pipelines. I thrive on bridging the gap between cutting-edge AI research and real-world business applications.
Currently leading AI initiatives at Rivercity Innovations, where I architect production multi-agent AI systems — combining open-source and commercial LLMs with real-time anomaly detection and predictive analytics — for enterprise IoT monitoring.
Architecting a production multi-agent conversational AI system (LangGraph) combining self-hosted open-source LLMs (vLLM) and commercial LLM APIs to deliver natural-language analytics over structured and unstructured operational data. Built a hybrid RAG and Text-to-SQL pipeline, and develop traditional ML and deep learning models for anomaly detection and forecasting.
Contributed to the Agency's internal generative AI frameworks and LLM tooling, helped establish AI deployment guidelines, and provided technical infrastructure support across ArcGIS, QGIS, and MSSQL environments. Presented AI concepts to 230+ non-technical stakeholders.
Conducted research on Distributed AI for 5G Networks. Applied federated learning to distributed, resource-constrained edge systems. Published 3 first-author papers in IEEE conferences and journals.
Enterprise conversational AI system combining RAG and Text-to-SQL pipelines with open-source (vLLM) and commercial LLMs, giving users natural-language access to real-time operational data.
Developed generative models to address data imbalance in EEG abnormality classification using a VAE-based data augmentation technique.
3D U-Net model with hyperparameter optimization for brain tumor segmentation on MRI imaging data. Graduation project, awarded Distinction.
Research on peer-coordinated sequential split learning for intelligent traffic analysis in mmWave 5G networks, and communication-efficient federated learning for UAV-IoT systems.
Comparative ML pipeline evaluating regression-based, tree-based, and RNN-based models for permanent magnet synchronous motor temperature prediction.
I'm always interested in discussing AI/ML innovations, research collaborations, and exciting opportunities. Feel free to reach out!