Mark Ogbodo.
I build the circuits AI thinks with — from neuromorphic hardware to production LLM and MLOps systems that hold up under real load.
My path runs from Nigeria to Japan to the U.S.: an MSc at the African University of Science and Technology, a Ph.D. in neuromorphic computing at the University of Aizu — where a neuro-inspired hardware architecture I co-developed led to a filed patent — and an MBA and Graduate Certificate in Business Analytics at Auburn University.
Along the way I've taught graduate courses on Network on Chip, co-supervised Ph.D. and Master's students, and led AI teams shipping computer vision, forecasting, and now agentic LLM systems into production. The throughline is the same at every layer: getting a signal from research to something reliable, whether it moves through silicon or a deployment pipeline.
Career signal trace
AI Solutions & Strategy Lead
- Engineered and implemented production MLOps pipelines automating continuous integration, training, and delivery of models using Python, Git, and containerization.
- Developed the technical infrastructure and deployment strategy for AI applications — scalable, fault-tolerant microservices on AWS integrating SQL/NoSQL data.
- Engineered LLM-driven systems and agentic AI workflows with LangChain and prompt engineering to orchestrate complex, multi-step reasoning applications.
Artificial Intelligence Tech Lead
- Developed and deployed computer-vision deep learning models for AI-based diagnostic applications using TensorFlow and PyTorch.
- Co-developed a multivariate sales forecasting system across 500+ SKUs, reaching 80% accuracy and enabling just-in-time procurement that cut inventory holding costs 15%.
- Led cross-functional collaboration with software engineers to integrate AI solutions into scalable business applications.
Artificial Intelligence Engineer
- Engineered a reconfigurable hardware-accelerated AI system for electric-vehicle energy management, prototyped on a Xilinx FPGA (C, Vivado) and a custom ASIC (Verilog HDL).
- Transitioned research into enterprise AI deployments, reducing model inference latency by 30%.
- Authored technical documentation that enabled easy project extension and scaling.
Graduate Assistant
- Created lecture notes and presentations for Business Analytics coursework.
- Proctored exams and evaluated student work against established rubrics.
Visiting Researcher / Instructor
- Taught a graduate-level course on Network on Chip.
- Co-supervised three Master's students and one Ph.D. student in deep learning and neuromorphic computing.
- Developed a prosthetic-hand control system that reached 87% performance relative to a conventional system.
Research Assistant
- Co-developed a novel neuro-inspired hardware architecture for deep learning (Python, MATLAB, Verilog, Cadence), resulting in a filed patent.
- Produced surveys, figures, and presentations communicating research methods and results.
- Published and reviewed journal and conference articles.
Foundations
Selected work
Co-developed a novel neuromorphic hardware architecture for deep learning using Python, MATLAB, Verilog, and Cadence — the work behind a filed patent.
Built a deep-learning control system for a prosthetic hand that matched 87% of the performance of a conventional control approach.
Reconfigurable, hardware-accelerated AI system for EV energy management — prototyped on a Xilinx FPGA and taped out as a custom ASIC in Verilog HDL.
Forecasting system across 500+ SKUs at 80% accuracy, enabling just-in-time procurement and a 15% cut in inventory holding costs.
Open source
Papers & writing
Recognition & service
Toolkit
Let's build something that runs.
Open to conversations on neuromorphic systems, applied AI engineering, and roles or collaborations where research needs to become a working product.