I like AI systems that help people read, understand, and organize scientific work. I am also interested in physics-informed machine learning and dynamical systems.

I hold an M.Sc. in Electrical Engineering from TU Hamburg. My thesis work focused on coupled port-Hamiltonian systems on graphs and neural-network-based modeling.

I currently work in a small B2B IT hardware company, where I handle IT operations, build internal Python and Microsoft Excel-based tools, automate business workflows, deploy internal systems, and improve operational processes with data.

Today, I build systems around retrieval-augmented generation (RAG), while continuing to study scientific machine learning, control systems, and dynamical systems.

A self-hosted research workspace for literature discovery, source-linked RAG, PDF annotation, Zotero workflows, spaced repetition, and project management.

JARVIS is local-first by default: it can run model inference through Ollama on the user's own hardware, while also supporting optional cloud models through LiteLLM for users who need more compute.

• Repo: https://github.com/FFidan/Jarvis-RD-Assistant
• Docs: https://ffidan.github.io/Jarvis-RD-Assistant/

• Dynamical systems: control theory, system identification, port-Hamiltonian systems, structure-preserving neural modeling, and neural differential equations.
• Applied automation: IT operations, internal tools, business-process automation, inventory analysis, document workflows, and practical digitalization.
• Engineering domains: physics-informed machine learning, electrical energy systems, smart grids, medical technology, aviation and aerospace, automotive, and autonomous systems.
• AI research tooling: retrieval-augmented generation (RAG), evidence extraction, and model-assisted literature workflows.