Applied AGENTIC AI, Machine Learning, and Data science

This project demonstrates a hierarchy of LLM agents that analyze energy microgrids by bootstrapping from a semantic model (Project Haystack + BRICK) rather than relying on bespoke hard-coded logic.

https://github.com/kgrajski/agentic_bems_paper

This project takes on the challenge of neural signal (ECoG) feature engineering to support long-term stable brain-computer interface (BCI) performance.

https://github.com/kgrajski/nvad_paper_scirpts

This project explores the frontiers of brain–computer interface (BCI) systems for decoding speech directly from neural signals recorded from implanted multi-electrode arrays.

https://github.com/kgrajski/speechBCI

This project implements an Agentic AI that monitors neuroscience sources daily, scores relevance with LLMs, and produces intelligence briefings here and here. It is deployed to OpenClaw and costs about a penny per run.

https://github.com/kgrajski/neurotech_newshound

This project operates at the leading edge of AI and Data Science for financial trading. An AI Agent orchestrates quantitative methods to generate candidate trades and applies LLM-based reasoning to evaluate and to advise.

https://github.com/kgrajski/trading_etf

This MVP R&D workbench applies AI/ML to model ERCOT real-time price vs day-ahead price (DART) dynamics. The project leverages API-based access to data from ERCOT historical and real-time data resources.

https://github.com/kgrajski/ercot_dart

This project is focused on multi-modal structural MRI and resting-state functional MRI (rsfMRI) biomarkers of Alzheimer’s Disease progression.

https://github.com/kgrajski/ADNI_Studies