Glotech.io
AI architecture
You can see which document each answer came from. We build RAG pipelines on pgvector and Pinecone and design agent orchestration with LangGraph. Guardrails keep answers tied to the source documents.
We build AI systems that keep
running in production
Technologies we use
Kubeflow
Google
Cloud
Terraform
Prometheus
Grafana
Docker
Python
Kubernetes
AWSAI and MLOps
MLOps engineering
Every trained model can be traced to the data and parameters behind it. We deploy training and serving pipelines on Kubeflow, define the infrastructure in Terraform and track experiments in MLflow. Any environment can be rebuilt exactly from code.
Platform operations
Each team sees what its own workloads cost. We run Kubernetes clusters on EKS and GKE and split costs per team and per service.
Cloud infrastructure and operations
High availability
The service keeps running when a single component fails. We set up multi-AZ deployments, automated failover with Route 53 health checks and self-healing pods.
Autoscaling
Clusters grow and shrink with real load, so cloud spend follows actual use. We set up pod autoscaling with HPA and node autoscaling with Karpenter.
AWS architecture
Workloads with strict compliance and audit requirements run in one AWS architecture. We use SageMaker for model training, Lambda for event-driven workflows, S3 for the data lake and CloudWatch for monitoring and alerts.
DevOps and observability
Your team can see what is running at any time. We set up CI/CD with GitHub Actions and ArgoCD, monitoring with Prometheus and Grafana, and alerting in PagerDuty.
AI system integration
If a model fails, the application falls back instead of breaking. We connect ML models to your applications through clearly defined APIs, with structured logging and circuit breakers.
Security and compliance
These controls go into the deployment pipeline on day one, so nobody has to add them later. We set up IAM policies, VPC isolation, KMS encryption at rest and audit logging for SOC 2 and GDPR.
Computer vision
We build computer vision systems that run in production, from the first dataset to deployment. We design each one for the place it will work: the lighting, the camera angles and the hardware on site.
Production-ready vision systems
Every prediction can be traced back to its input. We build inference pipelines with ONNX Runtime and TensorRT and serve them through FastAPI endpoints with health checks and logging.
Data-centric development
We review the data before we touch the model, and label it in Label Studio. In our experience, careful review improves accuracy more than moving to a bigger model.
Training, deployment and retraining
We train RT-DETR and custom detection models in reproducible experiments and release them through canary rollouts. Feedback from production feeds the next round of training.
Edge and cloud deployment
When latency matters, we deploy vision models to edge devices such as Jetson and Coral. Heavier workloads run on GPU clusters in the cloud. We measure latency, throughput and cost before anything goes live.
Dataset versioning
When a model gets worse, you can trace the cause to the exact data change. We version datasets with DVC and Git LFS, so training runs can be compared side by side.
Monitoring, drift and human review
You catch a drop in accuracy when real-world data changes. We track prediction confidence in production, flag data drift with Evidently and send low-confidence samples to a person for review.
Kaya
Kaya is Glotech's own product. It reads the tower light on each machine through the cameras your factory already has, and shows how many hours each machine actually ran.
From light to hours
Kaya reads each machine's tower light as green, amber, red or off. It turns that into hours per machine, per day and per shift. When the light is amber and nobody is at the machine, that time is shown separately and not counted.
Nothing is fitted to the machines
Kaya uses the cameras already on the factory floor. You don't add sensors, wiring or a PLC connection. If your camera recorder (NVR/DVR) can only be reached from inside the factory, one small computer sits on your network. It touches no machine. Make and age don't matter. Kaya works on any machine with a tower light: CNC machines, presses, injection moulding machines, packaging lines.
Every hour has a camera frame
Open any hour in a report to see the camera frames it was measured from, the light colour Kaya read and how sure it was. If the camera could not see the light, that time is marked as no data, never as zero.
Frequently asked questions
What is Glotech?
Glotech is an AI infrastructure and MLOps consultancy. We take AI systems from development into production.
Do you only work with AWS?
No. AWS is a large part of our work, but we also work on Google Cloud. Most of what we build runs on open-source tools that work on either cloud.
Is Kaya a consultancy project?
No. Kaya is a product Glotech built to measure how long machines run. It is sold separately from our consultancy work.