I’m an AI Platform Engineer and AI Solution Architect with a background in Data Science, based in Vienna. I enjoy working at the intersection of technology and business, understanding complex challenges and finding practical ways to solve them.
In my most recent role, I spent nearly four years at Hewlett Packard Enterprise in Advisory & Professional Services, working with clients across the DACH region. My work ranged from understanding business needs and designing technical solutions to building, deploying, and operating AI and data systems in production.
I’m naturally curious and enjoy both the technical side of engineering and working with people from different backgrounds. I believe the best solutions come from combining solid engineering, a product mindset, and genuine collaboration, with a focus on creating lasting value for the people and organizations using them.
I design and build AI applications and platforms, specialising in generative and agentic AI. I take solutions from concept to production, bringing the engineering needed to integrate AI into business operations and keep it running reliably.
02
Data Science
I use data and machine learning to help organisations plan ahead, detect problems earlier and understand what drives performance. I turn analytical work into models and findings people can use to guide their decisions.
03
Technology Consulting
I help organisations decide where AI and data are worth investing in and what it will take to deliver. I bring hands-on engineering experience to those decisions and work closely with clients to put them into practice.
Most recent experience
Where my technical depth became practical delivery.
Advisory & Professional Services · DACH Region
Hewlett Packard Enterprise
Data Scientist · Technology Consultant
At HPE, I supported clients across the DACH region as part of Advisory & Professional Services. Despite my role title, my work covered AI engineering and infrastructure, DevOps, data science, platform and data engineering, and the deployment and operation of ML and LLM systems for clients in automotive, the public sector, construction, real estate, industrial distribution and sports organisations.
Delivery roles across my tenure
01
For a state-owned real estate company’s digital AI avatar, I built and operated monitoring for an on-premises, multi-tenant Kubernetes platform. NVIDIA NIM and vLLM served the language models, while Milvus supported knowledge retrieval. Prometheus and NVIDIA DCGM Exporter captured service and GPU metrics; Loki collected logs through Promtail / Grafana Alloy. I deployed the monitoring components through Helm and versioned their configuration in Git.
I built Grafana dashboards and configured PromQL alerts with Alertmanager notifications to track response times, token throughput, errors, GPU usage and service-level targets. Monitoring covered time to first token, Milvus and retrieval-augmented generation (RAG) pipelines, with separate visibility for each tenant. I integrated it with the team’s role-based access controls, secrets management and network policies, and recommended GPU allocation and scaling changes based on performance analysis.
To improve answer quality, I tuned semantic retrieval in Milvus, applied NVIDIA NeMo Guardrails to constrain topics and responses, and evaluated model, prompt and RAG variants with MLflow. I also co-developed and integrated containerised LangChain and LangGraph agents and RAG components into production workflows.
02
For a global premium automotive manufacturer, I helped stabilise HPE Ezmeral Runtime Environment, a Kubernetes platform for production analytics and AI. I developed and automated a certificate-rotation procedure designed to keep workloads running without downtime.
After validating the procedure in a non-production environment, I documented the operating steps and rollback procedure and implemented it in production. I also configured identity integration, storage and platform health checks.
03
Across HPE projects, I built and maintained delivery pipelines with GitHub Actions and Jenkins to automate testing, container image builds and Kubernetes deployments. I packaged services and platform components as Helm charts with configuration for each environment.
Using Argo CD, I kept deployed workloads aligned with configurations stored in Git. This GitOps approach made deployment changes traceable and supported rollbacks through version history.
04
I co-developed a Python forecasting service for an internal HPE product, covering batch and streaming data ingestion, model training, deployment and prediction. I ran training, experiment tracking and model versioning in the AWS-based HPE Machine Learning Development Environment (Determined AI), using data from Amazon S3.
I deployed the containerised service on Kubernetes and exposed forecasts through a FastAPI REST interface for HPE Private Cloud AI, LLM applications and enterprise systems. I ran unit and integration tests with pytest, participated in code reviews and contributed to the design of Apache Flink stream processing for time-series data.
05
For a global automotive manufacturer, I built a Python transformation engine that processed data received through Apache Kafka for analytics in PostgreSQL and Neo4j. I containerised the services with Docker and administered both databases.
I co-designed and developed a Python similarity-search algorithm to improve metadata retrieval, with client-specific evaluation based on Jaccard similarity and Pearson and Spearman correlations. I also applied graph analysis and supervised classification to metadata, using Neo4j Graph Data Science for similarity, centrality and pathfinding. I explored the data and validated machine-learning results.
06
I developed and scheduled Apache Airflow workflows for a professional sports organisation’s lottery-based fan-engagement platform. I defined the data-processing steps and their dependencies, with retries and monitoring to support reliable, auditable execution.
07
For an international trading and distribution company, I wrote scheduled Python scripts to automate file retention, log rotation and server-side data lifecycle tasks. I also supported server upgrades, certificate renewals and runtime migrations as part of ongoing infrastructure maintenance.
08
For a national construction company, I conducted data-architecture workshops using HPE’s Edge-to-Cloud Adoption Framework (E2CAF). I mapped data capabilities, access arrangements and role-based permissions to support architecture and governance discussions. I also supported the establishment of a Transformation Business Office for governance and decision-making.
09
Through the HPE & NVIDIA AI Development Initiative, I trained in AI infrastructure operations, generative AI, large language models and multimodal AI. Additional training covered HPE Private Cloud AI and model-deployment practices in Azure ML.
During an LLM hackathon, I designed and built an autonomous agent with LangChain. I also delivered technical presentations on AI-based aerial image analysis and HPE’s generative AI portfolio, and took part in HPE DeliverAI Sessions.
