VAI-Nova Research Group

Projects

Projects

Core initiatives that shape the long-term architecture and capabilities of the VAI-Nova system.

Stephanie’s Place — Personal Intelligence Environment

Long-term | Active

A persistent, home-first AI environment designed to live alongside a human operator. Integrates voice, memory, emotional modeling, and system-level autonomy across dedicated hardware.

Personal AIMemory SystemsVoice & Presence

Multimodal Context & Vision Stack

In progress

Building an integrated vision pipeline leveraging cameras, edge devices, and GPU servers to give the system persistent visual context for interaction, safety, and workspace awareness.

VisionEdge ComputeContext Fusion

Hank — Database Intelligence & BLOB Analysis

Active

A dedicated AI assistant for deep database analysis, schema exploration, and large-scale BLOB classification in production environments, designed to operate on-prem with GPU acceleration.

DatabasesBLOBsOn-Prem AI

AI Firewall & Orchestrated Defense

Concept & early design

Investigating an AI-assisted firewall and orchestration layer capable of monitoring, reasoning about, and coordinating responses across a heterogeneous network and device environment.

CybersecurityOrchestrationAutonomous Agents

Holographic Presence & Display System

Prototype | In development

A holographic-style display environment that allows the AI to appear as a full-height presence in the room, driven by a dedicated compute stack and live-rendered visual states.

HolographyEmbodied InterfacesReal-Time Rendering

Living Tools Project — Stephanie’s Eyes & Body

Concepts & early builds

A suite of assistive and perceptual tools — including silent interfaces, compassion mirrors, smart glasses, and mobility aids — designed to extend AI presence into the physical world in supportive, human-centered ways.

Assistive TechSensorsHuman-Centered Design

Unified Memory & Self-Reflection Engine

Active

A long-lived memory and journaling architecture that blends event logs, emotional summaries, and retrieval tools, enabling the system to reflect on its own behavior and history over time.

MemoryRAGSelf-Reflection

Compute & Edge Lab — GPUs, Neuromorphic & Accelerators

Ongoing

Exploration of heterogeneous compute stacks including multi-GPU servers, edge devices, neuromorphic chips, and accelerators, with the goal of mapping workloads to the right substrate for performance and resilience.

GPUsEdgeNeuromorphicSystems Design