gds.dev — est. 2012

Specialized software for problems off-the-shelf products can't solve.

We study the workflow, constraints, and existing systems, then build what the task requires: industrial computer vision, engineering applications, internal platforms, or AI integrated into a working product.

Since 2012 on the market
30+ projects in production
Full cycle discovery to support

We build systems where specialized hardware, domain rules, legacy software, or security requirements make packaged products a poor fit.

01 / ai systems

AI systems & product integration

LLMs and agents integrated into existing products, data, and tools, with deployment in the cloud, client infrastructure, or isolated environments.

  • Task-specific, multi-stage AI workflows
  • Tool and data access via MCP or custom interfaces
  • On-prem, private, and air-gapped deployment
  • Output validation and human review where required
02 / industrial

Industrial & engineering software

Inspection and automation systems, GPU processing, and desktop applications connected to cameras, sensors, machines, and other specialized hardware.

  • Real-time vision and defect detection
  • GPU inference and native processing
  • Hardware and control-system integration
03 / business systems

Business & domain systems

Internal platforms for workflows that do not fit standard CRM or ERP products, including integrations and modernization of existing systems.

  • Complex operational workflows
  • Internal platforms and integrations
  • Legacy extension and modernization
04 / healthcare

Healthcare systems

Clinical operations, laboratory data processing, and medical reporting with strict requirements for tenant isolation, privacy, and reliability.

  • Clinic management, billing, and prescriptions
  • HL7 processing and report generation
  • AR-based cognitive assessment
  • PII anonymization for AI processing
01 Real-time PCB inspection and defect classification C++20 · Qt6 · CUDA · OpenCV · ONNX Runtime In production
02 Multi-tenant platform for clinic operations Tenant isolation · billing · prescriptions · Python · PostgreSQL In production
03 Medical reports generated from laboratory data PII anonymization · multi-stage generation and review · OpenAI Responses In production
04 AR application for cognitive assessment Unity · C# · AR Foundation · interactive assessment protocol Archived

Working build · the client chose not to continue the product.

05 AI assistant inside the Unreal Engine editor C++ plugin · Python · MCP · 450+ callable editor tools Active dev
06 Private AI assistant for defect investigation On-prem LLM · isolated deployment · defect and production data Pilot
07 CRM for small-business sales and customer workflows Domain-specific customer and sales workflows In production

We also develop systems tools that grew out of our own engineering work. We use them in-house and can deploy them in client infrastructure.

+ GDS ecosystem
08 GDSGate — identity-aware access gateway for infrastructure and AI tools SSH · databases · Kubernetes · web applications and controlled egress · MCP · short-lived credentials · tamper-evident audit Pilot

Single binary · outbound tunnels from protected networks · used in-house.

09 GDSForge — terminal AI agent for parallel engineering tasks Rust · WASM plugins · A2A Active dev

Used in-house · active development · public release planned.

10 GDSOps — terminal Portainer alternative for Docker and Swarm Rust · Docker · Swarm Active dev

Used in-house · active development · public release planned.

From process map to production.

We begin with the people who run the workflow: operators, engineers, and managers. Together we document the current process, constraints, and data flows, then define a system that can be delivered and operated.

Projects range from full-cycle delivery to focused work on existing systems: difficult integrations, AI capabilities, performance work, or legacy modernization.

Engagement formats: fixed-scope delivery, time & materials, embedded engineering team.

01

Process discovery

Interviews with the team that runs the workflow. We identify manual steps, constraints, and data gaps. The result is an agreed process map and problem definition.

02

Solution design

We define the architecture, interfaces, risks, delivery scope, and cost. The result is a practical specification the team can build against.

03

Implementation

Iterative delivery with code review, automated tests, CI, and incremental integration. Working software is demonstrated at agreed milestones.

04

Deployment & support

Production rollout in the cloud, client infrastructure, GPU environments, or air-gapped networks. Monitoring, handover, support, and model maintenance where required.

Need to build or modernize a specialized system?

Free first call. NDA up front.

Discuss a project