At CyberRant, I engineer Rant AI and production LLM capabilities that power the company’s core AI products. I also advance private foundational-model capabilities through fine-tuning, evaluation workflows, and inference infrastructure.
I built an AI Developer CLI and bootstrap system to automate environment provisioning, dependency validation, runtime detection, and configuration. I also designed and shipped reusable AI infrastructure for model integration, tool calling, context management, and workflow orchestration.
On Humind, I built multi-agent workflows and Python/FastAPI services for agent execution, orchestration, and LLM-driven task routing. I integrated context-aware reasoning across specialized agents and refactored workflow services into modular, testable components.
I built a document-processing system that automated intake and classification, cutting manual processing effort by about 60%. My other projects include a FAISS-based RAG pipeline for multi-source question answering and a large-scale agricultural ML pipeline that achieved about 83% classification accuracy.
