01
About the laboratory
SENTIA Lab studies and builds computational systems, with research spanning cognitive architectures, artificial life, security engineering, and distributed computing. Our projects address different technical problems; the laboratory is not defined by a single agent framework, chatbot, or language-model interface.
Our work includes AELIA, a study of persistent digital entities and cognitive life; RiFT, a cross-domain security agent and MCP tooling platform developed by RiFT Research, a subordinate group within SENTIA; and Lattice, a distributed resource network with locally governed compute nodes. Our work emphasizes substantive research questions and rigorous systems engineering.
PARENT LABORATORYSENTIA LabResearch direction, coordination and affiliated projects
DIRECT LABORATORY PROJECTAELIAPersistent artificial cognition and artificial life
SUBORDINATE RESEARCH GROUPRiFT Research ↗Security research group within SENTIA Lab; develops a broad security agent and MCP platform
DIRECT LABORATORY PROJECTLatticeDistributed compute resources, networking and workload coordination
Research clarification. Work on simulated consciousness, emotion-like state, and digital identity belongs to AELIA's research program. It does not imply subjective awareness. RiFT and Lattice are separate security and distributed-systems efforts, not subsystems of an AI assistant.
02
Research directions
Four overlapping areas, each with its own research questions and engineering methods.
- 01
Computational cognition & artificial life
Architectures for continuing digital entities: autobiographical memory, evolving self/world/belief models, affect-like dynamics, internal motivation, social cognition, and development through experience.
- 02
Security research & controlled execution
Research spanning web and API assessment, source-code auditing, reverse engineering, binary analysis, fuzzing, original vulnerability discovery, and L2/L3 network security.
- 03
Distributed networks & resource systems
Cross-platform compute nodes, encrypted overlay connectivity, local ownership and consent, signed workload leases, resource allocation, and scheduling across heterogeneous machines.
- 04
Systems architecture & experimental methodology
Typed interfaces, explicit trust boundaries, auditable execution, state continuity, failure recovery, and evaluation that separates intended behavior from demonstrated results.
03
Research projects
Research projects and affiliated groups of the laboratory.
Core research project
Cognition & artificial life
Autonomous Embodied Life & Intelligence Architecture. A research architecture for an evolving digital entity: continuous identity and autobiographical memory, belief/world/self models, simulated affect and drives, social relationships, theory-of-mind hypotheses, and internally motivated behavior. Its aim is computational cognition and artificial life, not a general-purpose tool-using assistant.
RiFT Research
Subordinate research group
RiFT Research is a subordinate security research group within SENTIA Lab. Its principal project, RiFT, aims to become a modular security agent and MCP ecosystem across offensive security research: black-box and white-box web testing, source auditing, reverse engineering, binary and firmware analysis, fuzzing, vulnerability and zero-day research, and network penetration testing at L2/L3. The research spans multiple security disciplines without being limited to any one type of target.
Laboratory project
Distributed computing
A cross-platform distributed resource network for user-approved computing workloads. Lattice combines enrolled nodes, overlay networking, device identity, local resource policy, signed job leases, verified runtime artifacts, and coordinated scheduling. Peer-to-peer connectivity is one part of a broader compute infrastructure.
04
Open questions
- Can a digital entity develop continuity?
- How can memory, self-models, affect-like state and learning shape a persistent artificial life across time and model changes?
- How can security tools remain constrained?
- How do we preserve explicit authorization and inspectable evidence when tools execute across bridges, sandboxes and agent-facing interfaces?
- Who controls distributed resources?
- How can owners consent to specific remote workloads while retaining enforceable control over identity, scheduling and resource budgets?
- What counts as experimental evidence?
- How do we distinguish an architectural proposal, a passing local test, and behavior verified in a real operating environment?