Moving beyond one-time login: passive signals, step-up challenges, and risk-aware sessions that re-verify users and devices as context changes—without grinding productivity to a halt.
A research line on self-sustaining companies: agent fleets that own routine decisions, closed-loop monitoring, and recovery—so creative work replaces repetitive labor at scale.
Grounded copilots for ERP, security, and operations: retrieval over private data, tool use with audit trails, and interfaces that explain not just answers but provenance and confidence.
Modeling how AI-augmented adversaries probe systems: forecasting weak paths, simulating lateral movement, and prioritizing hardening before incidents materialize.
Systems that watch their own obligations—data residency, access policies, and audit artifacts—continuously reconciling drift and generating evidence for regulators and boards.
Beyond rigid RPA: models that read unstructured inputs, negotiate exceptions, and coordinate with legacy systems—bridging the gap between tribal knowledge and executable workflows.
Prototypes are stress-tested against real enterprise constraints—latency, sovereignty, and auditability—so findings graduate into products like Dora ERP and our cyber practice, not slide decks alone.
Minimization, on-prem and VPC options, and encryption paths evaluated from day zero.
Joint programs with universities and select partners; publications where it accelerates the field.
Assistant and autonomy work includes misuse testing and operational guardrails.
Engineering near customers and global collaborators keeps research grounded in deployable software.
Whether you lead a lab, represent industry, or want to contribute as a volunteer, tell us how you’d like to collaborate. We’ll route you to the right team.
Thank you
Someone from Teamjaaf will contact you soon.