Mark Bain is the founder and chief executive of AIUS Technologies, which develops autonomous enterprise data agents capable of producing reproducible analyses and preserving the reasoning behind their work. His central technical interest is giving automated systems durable, connected memory of the people, data, decisions, and actions shaping their behavior.
Bain studied economics, information systems, strategic management, and marketing at the Warsaw School of Economics and attended a Stanford computer science summer session in 2014. He co-founded the education platform PERK, helped build technical-training company MrCertified, and served as an interim investment director at TYR TFI. At MrCertified, he managed engineers and instructors and rebuilt a cybersecurity training environment for teams participating in NATO Locked Shields exercises.
Before founding AIUS in 2024, Bain developed Aize, an assistant for querying documents, video, and audio. He subsequently created AIUS Memory, an open-source framework for long-term agent memory that separates episodic, entity, working, short-term, and long-term information across configurable storage systems.
At AI Engineer World’s Fair 2025, Bain led a workshop on graph-based agent memory, bringing together specialists from Cognee, Neo4j, and Zep while introducing his own ideas about evaluation and cybersecurity.
- Causal context in knowledge graphs. Bain argues that agents need relationships between events, decisions, and actions—not simply retrieved facts—to reconstruct context and make automated behavior easier to inspect.
- GraphRAG chat arena. His prototype evaluates memory systems through evolving agent interactions, comparing how integrations including Neo4j, Graphiti, Cognee, and Mem0 write, retrieve, and update connected information.
- Agentic firewalls. Drawing on cybersecurity work involving multiple shells and network environments, Bain envisions security controls that track commands, users, machines, sessions, and automated actions as episodic history.
- Deterministic context graphs. AIUS applies these principles to enterprise analytics, linking datasets, analytical workflows, decision traces, and reproducible outputs such as notebooks, scripts, models, and visualizations.
