Senior Freelance AI Engineer | LLM-Agents (Text & Voice), RAG, MCP, AWS

7/2024 – 10/2024

E-Commerce / Retail. Optimized the company's Data Lakehouse architecture built on Airflow, Spark and Dremio, improving query performance and reducing operational costs. Transformed infrastructure management with Terraform-based automation, minimizing deployment errors. Enhanced Kubernetes cluster stability through advanced monitoring and autoscaling, decreasing system downtime. Redesigned ETL pipelines for increased reliability, ensuring consistent completion of the core data flows supporting retail analytics.

Technologies:
• Kubernetes (AWS EKS, Helm, Grafana, Prometheus, Autoscaling)
• Infrastructure-as-Code (Terraform, AWS CDK, GitOps, CI/CD)
• ETL Pipelines (Airflow, EMR, Spark, PySpark, Python, Delta Lake)
• Observability (CloudWatch, Elastic Stack, Grafana, AlertManager)

Apache Spark, ETL, DevOps, Amazon Web Services (AWS), Kubernetes

4/2023 – 6/2024

Pharmaceutical. Developed an analytics platform for clinical trials that reduced operational decision time and improved trial efficiency. Engineered scalable Python microservices that processed complex trial data with high reliability. Implemented AWS CDK infrastructure that automated deployment tasks and optimized cloud resources. Created efficient ETL pipelines that transformed raw clinical data into insights for trial teams. Trained and deployed specialized LLMs that answered complex clinical queries, supporting critical decision-making processes.

Technologies:
• Large Language Models (Fine-Tuning, LangChain, Azure OpenAI)
• Python Backend (FastAPI, OpenAPI, Pydantic, Open-/ElasticSearch)
• Infrastructure-as-Code (AWS CDK, TypeScript, CloudFormation)
• ETL Pipelines (AWS Glue, Step Functions, Pandas, PySpark, Airflow)

ETL, Langchain, Large Language Models, Python, Microsoft Azure

3/2022 – 3/2023

Publishing / Media. Engineered and deployed the company-wide identity and data management platform that unified access across multiple business units and millions of users. Architected event-driven microservices with Go that processed high-volume transactions with low latency. Revamped data architecture using Databricks and Delta Lake, improving analytics performance. Implemented streaming data pipelines that reduced data processing time, enabling near real-time decision making for digital publishing operations.

Technologies:
• Event-Driven Architecture (Node.js, Async, Kafka, SQS, GraphQL)
• Cloud Infrastructure (EKS, ArgoCD, Lambda, Jenkins, CloudFront)
• Data Storage (S3, DynamoDB, RDS, ElastiCache, MongoDB, PostgreSQL)
• ETL & Data Engineering (Databricks, PySpark, Python, Airflow)

Apache Kafka, Databricks, Go (Programmiersprache), Amazon Web Services (AWS), Kubernetes

8/2021 – 2/2022

Automotive. Designed and implemented cloud-native infrastructure that scaled to support connected vehicles with high availability. Built real-time data processing services that reduced analytics latency, enabling timely customer insights. Developed machine learning pipelines that improved customer preference prediction, driving personalized experiences. Implemented infrastructure-as-code practices that accelerated deployment cycles, supporting rapid innovation for digital services under the SAFe ART framework.

Technologies:
• AWS Services (Glue, Athena, Step Functions, Lambda, SageMaker)
• Infrastructure-as-Code (CDK, Terraform, CloudFormation, Python)
• Microservices (Python, Java Spring Boot, REST, API Gateway)
• Containerization (Kubernetes, EKS, ECS, Fargate, Docker, ECR)
• Data Engineering (PySpark, EMR, Glue, Athena)

Docker, Machine Learning, Python, Amazon Web Services (AWS), Kubernetes

3/2021 – 7/2021

Energy / Climate (via Netlight). Architected a full-stack IoT data platform that processed environmental metrics from numerous sensors with high reliability. Implemented real-time analytics of time-series data using Azure Databricks that improved carbon capture optimization processes. Developed a low-code annotation interface that enhanced data labeling efficiency, accelerating machine learning model development. Created automated data pipelines that enabled precise measurement of carbon capture performance, supporting the company's climate impact goals.

Technologies:
• Microsoft Azure (Functions, Data Lake, Event Hub, Key Vault, Entra)
• Data Storage (PostgreSQL, Time-Series, Blob Storage, Data Lake)
• Data Processing (Databricks, PySpark, Python, Delta Lake)
• IoT (Azure IoT Hub, IoT Edge, Device Provisioning Service)

Databricks, Postgresql, Python, Microsoft Azure, Internet of Things (IoT)