Certified Research Organization · ID 2025/10547-E1230Certified Research Org · 2025/10547-E1230[ 01 — Software house · Est. 2018 ]

We engineerintelligent agents.

QUES is a Bratislava-based studio building AI-native products, autonomous agents, and the infrastructure they run on. From first principles to production.

What we buildAI AgentsLocal AIAI InfrastructureWeb · Mobile · IoTHealthtechDefense / ISTAR
AI agents
Local AI
AI infrastructure
Multi-agent systems
Edge inference
RAG pipelines
LLM ops
Data platforms
Cloud architecture
Healthtech
Defense / ISTAR
Web · Mobile · IoT
AI agents
Local AI
AI infrastructure
Multi-agent systems
Edge inference
RAG pipelines
LLM ops
Data platforms
Cloud architecture
Healthtech
Defense / ISTAR
Web · Mobile · IoT
[ 02 — AI Lab ]

Three pillars.
One stack.

AI is not a feature we bolt on. It's the substrate we build with. We ship agents to production, run open models on your own hardware, and operate the infrastructure they need to be fast, cheap, and trustworthy.

01 / 03

AI Agents

Autonomous, tool-using, production-grade.

We design and ship multi-agent systems that plan, call tools, and operate inside your data — from customer-facing copilots to internal automation that retires entire workflows.

  • Multi-agent orchestration · planners + executors
  • Tool use, MCP, structured output, evals
  • Deterministic guardrails & human handoff
  • Memory, RAG, knowledge graphs
Talk to our AI team
02 / 03

Local AI

Inference where the data lives.

Sovereign, on-prem and on-device intelligence. We deploy open models on your hardware — from regulated healthcare environments to ruggedized edge devices in the field.

  • Llama / Qwen / Mistral on your hardware
  • Quantization, distillation, fine-tuning
  • Edge inference (NVIDIA Jetson, Apple Silicon)
  • Fully air-gapped deployments
Talk to our AI team
03 / 03

AI Infrastructure

The platform underneath the magic.

GPU clusters, inference gateways, vector stores, observability. We build the boring, hard parts that make AI products fast, cheap, and reliable at scale.

  • GPU orchestration · Kubernetes · Ray
  • Inference gateways & cost routing
  • Vector databases & data pipelines
  • Tracing, evals, drift monitoring
Talk to our AI team
[ 05 — Start something ]

Have an idea
that should think?

We take on a handful of partnerships each year. If your problem is interesting, your team is ambitious, and you care about craft — we'd love to talk.

Response time
< 24h
Engagement length
8 wk · 2 yr
Active partners
12+
Office
Bratislava · EU