At Technosec.io, I build and maintain a JupyterLab-based econometric and marketing-analytics platform across Python, React, and TypeScript. I've owned core frontend extensions, multi-step data-pipeline workflows, a React Flow graph editor, and a standalone app shell built around a plugin architecture.
I've led the platform's AI integration through Datanout, a notebook-native LLM assistant with LangGraph tool calling, human approval, MCP support, and cross-session memory. I also harden multi-tenant filesystem and persistence layers against path traversal, symlink attacks, SSRF, and environment-variable injection.
I'm an upstream contributor to JupyterLab and Jupyter-AI, with merged security, file-handling, and agent-stack improvements. My work also includes PanelBoom, encrypted image-generation tooling, FastAPI LLM inference services, and Next.js analytics dashboards.
