04.09.2026 aktualisiert
Premiumkunde
100 % verfügbarSenior Cloud & Gen AI Solutions Architect
Dresden, Deutschland
Weltweit
Über mich
Gen AI / KI Solutions Architect | Generative AI, RAG, Agentic AI, AI Governance | AWS Bedrock & Azure AI | Banking, Insurance, Pharma, Automotive | working languages EN/CZ
Skills
APIsKünstliche IntelligenzAmazon Web ServicesCloud-EngineeringAzure Machine LearningSolution ArchitectureCloud PlatformGenerative AIKubernetesCloud-MigrationTerraform
Senior Cloud & Gen AI Solutions Architect with 20+ years of delivery experience in banking, insurance, pharma, automotive, aerospace and fintech. I design production-grade Generative AI systems and the multi-cloud platforms they run on (AWS first, Azure, GCP/OCI) — from RAG and agentic Large Language Model (LLM) architectures to AI governance that keeps agents inside regulatory and corporate rules. Architecture-led, hands-on where it matters (PoCs, MVPs, IaC).
— RECENT HIGHLIGHTS —
· Policy-as-Code for AI agents (Kyndryl, banking): designed the integration of a policy platform into a bank's AWS environment and led its implementation — regulatory and internal norms (EU AI Act, DORA, GDPR) turned into machine-enforceable guardrails for LLM agents on Amazon Bedrock
· LLM decision automation for a call center on AWS: call recordings transcribed and interpreted into structured signals (intent, outcome, next step), combined with customer-data predictions and fed into automated downstream processes
· Pharma Gross-to-Net AI analytics platform: multi-agent Text-to-SQL framework (semantic layer, Bedrock AgentCore with LangGraph agents), Aurora + pgvector, EKS Auto Mode — greenfield architecture and MVP
· AWS Landing Zone and on-prem migration for the Botswana Stock Exchange; multi-tenant AI analytics with QuickSight for call-center insights
· Privacy-first automotive AI: fine-tuned open-weights Hugging Face models on SageMaker with OpenSearch vector RAG inside the VPC
· Multi-cloud migration program for a major EU banking group — AWS migration patterns, greenfield Azure platform, shared services and AI platform design (Raiffeisen Bank); earlier: group cloud platform and AI assistant for KBC/ČSOB
— CORE FOCUS —
AI & GenAI: Amazon Bedrock, AgentCore, Guardrails, SageMaker | Azure AI Foundry, Azure OpenAI | LangGraph, LangChain, LiteLLM Proxy, LangFuse, Hugging Face, MCP | RAG, agentic AI, Text-to-SQL, fine-tuning | vector stores: pgvector, OpenSearch, Qdrant
AI Governance: Policy-as-Code, AgentCore Policy (Cedar), audit trails, EU AI Act / DORA-aware guardrail design
Cloud Platforms: AWS (EKS, Aurora, landing zones, IAM/OIDC federation) | Azure (AKS, Cosmos DB, Entra ID) | GCP, OCI
Engineering: Terraform, Terragrunt, Kubernetes, GitHub Actions | Python, Java | PostgreSQL, Redis | REST APIs, OAuth2/OIDC
AWS Certified Solutions Architect – Associate | AWS Certified Machine Learning Engineer – Associate
Für DACH-Auftraggeber: KI-Architekt für Künstliche Intelligenz (KI), Generative KI, KI-Plattformen und KI-Governance auf AWS und Azure. Deutschsprachige Unterlagen lese ich, schriftliche Kommunikation auf Deutsch ist mit KI-Unterstützung möglich – Projekt- und Arbeitssprache ist Englisch.
Remote-first contract/B2B (part-time possible, up to 20% on-site in Europe). Working languages English (fluent) and Czech (native); open to English-speaking projects in DACH, Ireland and the EU. Based in the Czech Republic.
Erfahrung
Berufserfahrungmehr als 10 Jahre
Freelancer seit2004
Sprachen
DeutschGrundkenntnisseEnglischverhandlungssicherSpanischGrundkenntnisse
Projekthistorie
Policy-as-Code for AI Agents (Kyndryl, banking): Designed the integration of a Policy-as-Code platform into a bank's environment and led its implementation, turning regulatory and internal norms into machine-enforceable rules for AI agents in bank processes.
Cross-functional delivery with the bank's security, platform and process owners.
LLM-based Decision Workflow Automation (call center): Designed an AWS-based AI solution that automates a call-center decision workflow: call recordings are transcribed and interpreted by an LLM into structured signals (intent, outcome, next required step), combined with predictions based on customer data and interaction history, and turned into decisions that drive the downstream process automatically.
Scalable AI Sales Analytics
Led the end-to-end architecture and MVP implementation for a massive-scale analytics solution featuring an AI Text-to-SQL semantic layer, LLMProxy, AWS Bedrock, AgentCore on LangGraph, Aurora RDS with vector extensions, and EKS AutoMode.
Enterprise LLM Integration
Designed and validated a PoC for integrating Large Language Models with Confluence and Salesforce, utilizing AWS Knowledge Base and S3 Vector for advanced knowledge retrieval.
AWS RAG Architecture
Conducted expert design reviews and provided strategic consultation for Retrieval-Augmented Generation architectures.
Cloud Solutions Architect for the CMAP (Cloud Migration and Adoption Program) project that encompasses the migration of a wide range of applications from on-premise to AWS, Azure and Oracle Cloud, as well as the design and governance of shared platform components and hybrid solutions design across on-premise and multiple cloud environments.
Botswana Stock Exchange — Migration to AWS: Designed the AWS Landing Zone for the exchange and oversee its implementation and the migration of on-premise workloads to AWS.
AI Multi-tenant Analytics
Led the solution design and PoC implementation for integrating AWS QuickSight into a multi-tenant AI environment to drive call center performance insights.