Case Studies
Selected examples of enterprise platform and AI delivery work, focused on architecture, delivery, and operational outcomes.
AI Customer Support & RAG Platform — WineSpot.ai
AI-powered customer support platform delivering accurate, up-to-date responses using retrieval-augmented generation and cloud-native infrastructure.
Focus areas
- LLM and RAG platform engineering
- Near real-time knowledge ingestion
- Kubernetes-based microservices on Azure
- Firebase Realtime Database for instant events delivery
Enterprise LLM Platform Enablement — Mercedes-Benz
Design and delivery of internal LLM platform foundations supporting multiple enterprise use cases, including LLM-based voice assistant capabilities.
Focus areas
- LLM gateways and governance
- Secure access to multiple model providers
- Kubernetes, GitOps, and observability
- Content filtering and accounting for streamed content delivery
- Token-based rate-limiting for various consumers
- Resource based loadbalancing over different model backends
- PTU management for exclusive access to model resources from enterprise level applications
SDN & Open Access Network Platform — Deutsche Telekom
Cloud-native SDN platform based on OpenCORD and VOLTHA, supporting end-customer internet access services in telco-grade environments.
Focus areas
- Software-defined networking (SDN) control-plane platforms
- Open access network stacks using OpenCORD and VOLTHA
- Self-managed Kubernetes deployments for telco workloads
- Scalability and reliability under real end-customer traffic
- Automation, observability, and operational resilience
End-to-End RAG Platform for Enterprise Knowledge Access
A multi-tenant RAG platform enabling controlled document ingestion and AI-assisted knowledge access for different consumer groups, including customers, partners, and internal teams.
Focus areas
- Enterprise RAG platform architecture
- Document ingestion and processing pipelines
- Vector search and AI-assisted knowledge access
- Fine-grained access control and governance
Computer Vision Annotation Platform for Automated Checkout (CVAT)
Enterprise-grade computer vision platform enabling scalable data annotation and model iteration for automated checkout use cases, deployed on EU-based cloud infrastructure.
Focus areas
- Computer vision platforms and data annotation
- CVAT hardening for enterprise use
- Kubernetes-based delivery and operations
- Compliance, scalability, and partner access