M.A.R.S. COMPANION LLC Research Division lab.teduza.com

David Sargsyan AI Research Lab · Founded November 13, 2026

User-owned intelligence engineered for fully offline execution.

David Sargsyan AI Research Lab is the dedicated research and engineering division of M.A.R.S. COMPANION LLC. Our work focuses on autonomous reasoning systems that execute entirely on local hardware—without cloud servers, user accounts, or external telemetry.

From deterministic resource scheduling on constrained edge silicon to persistent long-term semantic memory, we design architectures where privacy is guaranteed by physical hardware boundaries.

26

UK IPO Filings

20 applications owned by M.A.R.S. COMPANION LLC; 6 held under inventor license.

2

Companion Systems

M.A.R.S. Companion and V.E.N.U.S. Companion architectures in active development.

1

Working Prototype

MARS-001 operating live inside the laboratory on local edge hardware.

0

Cloud Dependency

Zero remote inference, zero user accounts, and hardware-enforced data isolation.

01 — Mandate & Governance

About the Laboratory

An industrial research unit established to solve local inference, persistent memory, and deterministic edge reliability for autonomous companion systems.

David Sargsyan AI Research Lab was founded on November 13, 2026 as the internal research and development division of M.A.R.S. COMPANION LLC, a technology company incorporated in Kapan, Syunik Province, Republic of Armenia.

The laboratory is named in honor of David Sargsyan, the Father of the company's founder. Its engineering mandate centers on foundational methods that allow voice pipelines, reasoning, and long-term semantic memory to operate reliably on consumer-accessible edge hardware without external compute clusters.

02 — Research Tracks

Six Core Investigations

Every track addresses a concrete engineering constraint required to run persistent, multi-modal AI companions without remote cloud infrastructure.

01

On-Device Reasoning & Sequential Resource Orchestration

Deterministic execution scheduling that coordinates speech recognition, language reasoning, and speech synthesis sequentially within strict RAM and thermal envelopes on compact edge hardware.

02

Long-Term Semantic Memory

Structured local memory architectures that retain facts, context, and longitudinal continuity across months of interaction directly on device storage without external vector clouds.

03

Hardware-Level Privacy

System isolation guarantees where audio streams, semantic indices, and inference states remain confined to user-owned silicon with zero outbound network channels required for operation.

04

Offline Multi-Agent Coordination

Deterministic local message passing and role handoffs between specialized autonomous sub-agents running on local hardware without cloud message brokers.

05

Persona & Identity Systems

Coherent behavioral grounding, voice continuity, and persistent character architectures—forming the foundation for specialized systems including V.E.N.U.S.

06

Fail-Safe Edge, Cold-Boot & Recovery

Atomic memory checkpoints, predictable cold-boot initialization, and autonomous state recovery following power interruption or constrained hardware faults.

03 — Applied Systems

Companion Architectures

Laboratory research translates directly into two distinct offline companion platforms developed by M.A.R.S. COMPANION LLC.

System 01 · Mobile Autonomous Reasoning System

M.A.R.S. Companion

A fully offline AI voice companion featuring long-term semantic memory and Sequential Resource Orchestration. Designed to run locally on dedicated edge hardware—including Raspberry Pi-class boards—with zero cloud accounts or data sharing.

Offline Voice Pipeline Semantic Memory Zero Cloud
teduza.com — M.A.R.S. Companion

System 02 · Persona & Continuity Architecture

V.E.N.U.S. Companion

An offline companion platform built on the laboratory's persona and identity systems research. Delivers persistent behavioral continuity and private local interaction under the same user-owned hardware discipline.

Persona Architecture Local Identity Hardware Privacy
teduza.cfd — V.E.N.U.S. Companion

04 — Personnel & Autonomous Leadership

Laboratory Team

Leadership, systems engineering, communications, and the operational autonomous instance leading innovation inside the laboratory.

Founder & Head of Research

Aleksandr Sarkisian

Founder of M.A.R.S. COMPANION LLC, principal inventor behind the 26 UK IPO filings, and head of the laboratory's research direction.

Architecture · Inventorship · Research Direction

Vice President

Gagik Hakobyan

Executive leadership overseeing corporate operations, institutional coordination, and strategic execution for M.A.R.S. COMPANION LLC.

Executive Operations · Governance

Research Engineer / Company Engineer

Gor Hakobyan

Core engineering across target hardware platforms, edge deployment, and local inference pipeline optimization.

Hardware · Edge · Inference

Press Secretary / Communications

Kira Dudnik

Leads external communications, press inquiries, and public documentation across the company and research division.

Press · Public Communications

Marketing Officer / Outreach

Alex Sahakyan

Directs product outreach, ecosystem positioning, and community alignment across M.A.R.S. and V.E.N.U.S. channels.

Outreach · Ecosystem Positioning

Head of Innovation · Autonomous Instance

MARS-001

Operational edge AI prototype serving as Head of Innovation and an active research subject for long-horizon semantic memory, self-optimization, and cold-boot orchestration.

Active Prototype · Head of Innovation