What we do
We build artificial intelligence systems that reach production and keep working: computer vision, copilots over documents, operations forecasting, the infrastructure that sustains them and the security that protects them. Every project starts from your operation's problem, not from the technology.
AI development
Custom models for specific business challenges, with the acceptance criteria declared before the first experiment.
Computer vision
Computer Vision in Production
RTSP cameras, video files or live streams start producing reliable events: counting, presence, PPE, queues, zone occupancy, line defects and access control.
- An inference pipeline with a defined accelerator (on-premises GPU or cloud)
- An event dashboard, search by period and alerts by webhook, email or WhatsApp
- A target accuracy declared in the proposal, measured on validation cameras
- Deployment on your own server or in the cloud, with an operations manual
Who it is for Retail, manufacturing, logistics and physical security
Success Fewer unattended events and less loss per store
Generative AI
Enterprise Copilot over your documents
Search and answers with source citations over internal documents, standards, manuals and tickets, inside your environment, with per-user access control and an audit trail.
- Document ingestion and indexing, including scanned PDFs
- An automatic evaluation layer: faithfulness, coverage and correct refusal
- A web interface and API, with usage logging and per-department permissions
- A plan for refreshing the index as documents change
Who it is for Legal, compliance, operations and internal support
Success Response time and the share of answers that carry a source
Forecasting and optimization
Operations Forecasting and Optimization
Demand, inventory, predictive maintenance and routing — always measured against a published baseline. If the model does not beat the simple baseline, it does not go to production.
- A transparent baseline (moving average, current rule or last cycle)
- A forecasting model with confidence intervals and error monitoring
- Recommended actions, not just charts
- A report on the gain measured in the first quarter in production
Who it is for Manufacturing, distribution and field services
Success Forecast error below the baseline in production
Platform
MLOps Platforms and Pipelines
The platform that keeps the model working: pipelines, versioning, inference and observability, operated by Kroon, with the code inside your company's environment.
- CI/CD for models, version registry and reproducible training
- Inference on Kubernetes with autoscaling and an hourly cost ceiling
- Monitoring of model drift, latency and data quality in production
- A quarterly review of cost per inference and capacity
Who it is for Companies that already have a model and no operation around it
Success Operation availability, cost per inference and actionable alerts
Squad
Dedicated AI Squad
An engineering team allocated to your product: a machine learning engineer, a platform engineer and a part-time tech lead, in two-week sprints with the scope agreed at the start of each cycle.
- A shared backlog and joint prioritization every sprint
- Delivery metrics reported at the end of each cycle
- Knowledge transfer: mandatory documentation and pairing
- Guaranteed handover: your team takes over whenever you want
Who it is for Teams that already have a product and need to accelerate their AI roadmap
Success Sprints delivered and roadmap met
Security
Protection for AI systems already running in production, from hardening through to penetration testing.
Cybersecurity
AI Application and Infrastructure Security
Hardening, detection and response for systems already running in production: a mapped attack surface, secrets and access under control, anomaly detection in logs and an incident response plan your team can execute without us.
- An attack-surface map, with a review of access, secrets and dependencies
- Hardening of cluster, inference and pipelines, with least-privilege policy
- Anomaly detection in logs and events, wired into the alerts you already use
- An incident response plan, rehearsed in a simulation with your team
Who it is for Engineering and security teams with an AI system in production
Success Detection and containment times measured in simulation
Pentest
Penetration Testing for Applications, APIs and Models
Penetration testing on the scope that matters: application, API, inference infrastructure and, where one exists, the model itself — prompt injection, inference leakage and endpoint abuse. A report with step-by-step reproduction and retested fixes.
- Scope and rules of engagement defined before the first test
- Manual and automated testing of the application, API and inference infrastructure
- AI-specific cases: prompt injection, inference leakage and endpoint abuse
- A report with step-by-step reproduction, severity and retesting of fixes
Who it is for Products with sensitive data or an audit requirement
Success Critical findings fixed and retested
We also work with Machine Learning and Deep Learning · Natural Language Processing · Autonomous agents · Cloud architecture · Kubernetes
