AI agent platform · 2025 to 2026

Deploying NVIDIA AIQ on customer-owned DGX Spark hardware

Deployed NVIDIA AIQ on NVIDIA DGX Spark and extended it with tools and agent capabilities, on customer-owned hardware.

NVIDIADGX SparkOn-premiseAgents
DGX Spark on-premise hardware
AIQ agent platform
On-prem deployment model

The short answer

We deployed NVIDIA AIQ on NVIDIA DGX Spark hardware and adapted it with tools and agent capabilities. The work was deployment and extension of an existing agent platform on customer-owned infrastructure, not the design of a retrieval system from scratch.

Context and role

ContextComputacenter Romania · customer project · September 2025 to July 2026
RoleDeployed and extended the platform as part of the delivery team.
CategoryAI agent platform

The problem

The customer wanted an agent platform running on hardware they owned, not on an external API. That changes the deployment problem: the hardware is fixed, the network boundary is real, and the team that operates it is the customer’s.

NVIDIA AIQ provides an agent platform, but the customer’s use case needed additional tools and agent behaviour beyond the default setup.

The platform had to be demonstrably running on the DGX Spark, with a path for the customer’s team to understand and operate it.

Approach

  • Deployed NVIDIA AIQ on NVIDIA DGX Spark hardware in the customer environment.
  • Adapted the platform with tools and agent capabilities for the customer’s use case.
  • Worked within the constraints of on-premise hardware and a controlled network boundary.
  • Handed the deployed setup to the team responsible for the environment.

What I did not claim

The boundary of a project is part of its evidence. These are the claims this work does not support.

  • Not a from-scratch RAG architecture. The work was deploying and extending an existing NVIDIA platform.
  • Not a production system with real users. There is no evidence of that in the project record.
  • LangChain was not used and is not claimed for this work.
  • No scale, uptime or performance numbers are claimed.

Results

  • NVIDIA AIQ ran on customer-owned DGX Spark hardware.
  • The platform was extended with tools and agent capabilities for the use case.
  • The project produced hands-on experience with on-premise agent deployment, which is rare in this price range.
  • The deployment stayed on infrastructure the customer controls.

Stack

  • NVIDIA DGX Spark
  • NVIDIA AIQ
  • On-premise deployment
  • Agent tools

What this means for you

The useful part of this project is not the artifact. It is the method: define the measurement first, keep the scope honest, and hand over something your team can run. That method is what transfers to a new problem.

If your situation looks similar, the related services below are the place to start. If it does not, send a short description and I will tell you whether I am the right person.

Boundary: This page describes deployment and adaptation work. It does not claim a production system, external users or a from-scratch retrieval architecture.

Written by

Marius Boroica is an independent AI automation engineer in Cluj-Napoca, Romania. He designs AI workflows, document pipelines and internal assistants, and connects them to the tools a team already uses. NVIDIA Certified Professional: Gen AI LLMs.

Page last updated: October 2026

Next step

Have a problem in the same shape?

A 30-minute call is enough to see whether the method transfers, and which engagement is the smallest useful one.

Prefer email? boroicamarius01@gmail.com