AI · machine learning · robotics
Two-time founding engineer. CTO at SpineSense and VP of Engineering at AcreAtlas.
I love modern AI, old-school machine learning, and robots—especially when the best answer mixes them.
How I work
I’m an engineer and technical leader who likes the full path from a rough research idea to software people can use. That means deciding where a model helps, where deterministic logic is safer, and how the whole product behaves under real constraints.
I started in mechanical engineering and robotics, moved through autonomy and applied machine learning, and now build health and agriculture products. I’m happiest when the work crosses perception, data, control, and product engineering.
Current work
I work across architecture, models, product, and engineering teams at two early-stage companies.
01 / Health
–present
CTO & Founding Engineer
Building patient-facing systems and clinical workflows across a Python/FastAPI and PostgreSQL backend, React Native/Expo and Next.js clients, with Azure OpenAI routed through LiteLLM. Deterministic red-flag and urgency screening runs alongside adaptive questioning.
02 / Agriculture
–present
VP of Engineering & Founding Engineer
Turning field imagery into weed detection and spray maps for farmers and researchers. The system combines YOLO/U-Net vision models, photogrammetry, orthomosaics, and geospatial processing through FastAPI, Next.js, Azure Batch, and PostgreSQL.
Technical range
Agentic workflows, model selection, Azure OpenAI, LiteLLM, multimodal PyTorch systems, and LLM evaluation.
Gradient boosting, random forests, and SVMs—practical tools alongside modern neural models.
ROS 2 perception, sensor interfaces, planning and control, and simulation-to-real workflows.
Python/FastAPI, TypeScript/Next.js, PostgreSQL, containers, and Azure infrastructure.
Earlier work
Boeing · applied ML & autonomy
At Boeing, I worked across autonomy, robotics, and applied machine learning, including ROS 2 perception, sensor interfaces, planning and control, and simulation-to-real workflows. Earlier autonomy work included tanker planning; later ML research included search-and-rescue, multimodal PyTorch systems, and LLM evaluation.
Robotics research
At Saint Louis University, I researched tactile and kinematic learning, vision-guided robotics, telepresence manipulators, and medical devices while completing an M.S. in Engineering with an AI focus. I also hold a B.S. in Mechanical Engineering from Southern Illinois University Edwardsville.
Selected public work
A few examples of the side work that keeps me learning.
Prototype · real-time translation
A TypeScript prototype for Even G2 smart glasses. A G2 client streams speech to an authenticated Azure relay, which checks the language and returns English translations with suggested replies. Spanish and Turkish are the initial target languages.
Integration · local AI interface
Connects Even G2 smart glasses to the upstream OpenClaw assistant through a Python WebSocket gateway with faster-whisper and a TypeScript/BLE client. Speech comes in through the glasses and streamed responses return to their display.
Experiment scaffold · bioinformatics
Python loaders, competition-metric evaluation, and submission validation for assigning Gene Ontology terms from amino-acid sequences in the CAFA 6 competition.
Earlier project · quantitative ML
A TensorFlow and pandas framework for defining features, prediction targets, and LSTM architectures for quantitative analysis. It grew from a Quantopian experiment and remains an early quantitative-ML playground.
Research & publications
Research across telepresence, medical devices, haptics, and electromechanics shaped how I think about intelligent systems.
Patent disclosure: Systems and Methods for Guiding Surgical Tools · PCT/US20/27625
Let’s talk
I’m always interested in thoughtful work at the edge of AI, classical machine learning, robotics, and useful products.