One deployment partner. Every place AI needs to run.

We design around the operating constraint, not a vendor catalogue. Each mission can combine cloud scale, private control, edge speed, and physical systems.

Direct answer

MH – Applied AI provides forward-deployed engineering for enterprise AI systems, agents and automation, private/local AI, cloud and hybrid architectures, robotics, and emerging personal-robot adoption.

Enterprise AI systems

Select, integrate, evaluate, govern, and operate AI around a measurable business outcome.

Agents and automation

Bounded AI agents and agentic coding with real tools, repository context, approvals, observability, and human control.

Private and local AI

Keep sensitive models, retrieval, and data paths inside infrastructure you control.

Cloud and hybrid AI

Use provider scale without surrendering architecture, portability, or cost discipline.

Robotics and embodied AI

Connect perception, edge inference, orchestration, and enterprise workflows to physical systems.

Personal Robotics Lab

Prepare for trusted personal robots through research, service design, integration, and responsible pilots.

Founder operating proof

From 24×7 robot operations to embodied AI.

Marcel’s robotics background predates today’s generative and embodied AI wave. He led engineering for the technical foundation around AutoStore systems in the Swisslog environment, supporting highly available operation, international on-site service, and 24×7 support.

The work connected warehouse management, conveyors, picking, real-time computing, monitoring, automation, incident response, and operational handover. That experience is why MH – Applied AI treats uptime, observability, human supervision, and knowledge transfer as part of the robotics system rather than aftercare.

The same skills context extends into the CNEXT AI Labs Misty project: the team Marcel co-leads connected Azure Speech, Azure OpenAI, automation, and live ERP data to an open social-robotics platform. This is CNEXT AI Labs team work, not a solo implementation.

AutoStore Technical foundation
International On-site service
Operations 24×7 availability
Misty II Embodied AI interface

Explore the robotics practice Read the Swisslog record Read the Misty record

Start with placement, not platform.

Where an AI workload runs determines latency, privacy, cost, resilience, and who can operate it. We make that placement decision before products harden into architecture.

A common engineering spine.

Outcome and risk framing
Data and identity boundaries
Model and tool evaluation
Integration and observability
Human approval and adoption
Operational transfer

Technology agnostic does not mean indecisive.

We make explicit recommendations, document trade-offs, and select the smallest credible stack. Independence means the decision belongs to the mission rather than a reseller relationship.

Bring us the mission, not a shopping list.

In the first conversation we map the operational outcome, constraints, deployment environment, and the shortest credible path to evidence.

Book an AI deployment consultation