Anton Rösler

Freelance Engineer

AI & Cloud Engineering

I build production-ready agentic AI systems: multi-agent architectures, LLM platforms and scalable cloud infrastructure.

AWS Azure Terraform Python Databricks Docker

About

I'm a freelance AI engineer specializing in agentic AI systems and production LLM applications. My focus: the full agentic runtime stack. From orchestration and context engineering to guardrails, observability and autonomous agents.

I'm currently engineering complex multi-agent architectures for enterprise AI platforms: orchestration, tool dispatch, human-in-the-loop approvals, durable execution, plus evaluation and tracing (Langfuse, promptfoo). Grounded in data engineering, MLOps and cloud, with Python, Terraform, AWS CDK and FastAPI on AWS, Azure and Databricks.

B.Sc. Engineering Business Information Systems (top grade)

Experience includes building large ML infrastructures at BOSCH.

Services

From initial consulting through MVP development to ongoing operations. I support projects end-to-end or step in where needed.

Software Engineering for AI/ML

Building production-ready AI services and applications. Python, APIs, microservices, event-driven architectures.

AI Agents & Platforms

Agentic AI, multi-agent systems, context engineering. Skill frameworks and enterprise integrations.

Cloud Engineering & Infrastructure

IaC with Terraform for multi-account/multi-region setups, networking, IAM, container platforms and cost optimization.

MLOps & CI/CD

Model registries, automated training/inference pipelines, GitHub Actions, canary releases, blue/green deployments.

Data Engineering & Lakehouse

Databricks / Delta Lake design, Spark/PySpark ETL/ELT, data cataloging, quality monitoring.

Observability & Operations

Logging/tracing, alerts & runbooks, SLO/SLI definitions, incident playbooks.

Selected Projects

BOSCH

Multi-tenant inference & data pipelines; Databricks lakehouse, Spark ETL.

Azure Databricks Spark Delta Lake

Cloud orchestration, scalable batch pipelines, custom AMIs/container images for automated 3D labeling.

AWS Batch Terraform Docker GPU/CUDA

Enterprise Client

AI platform: Secure infrastructure, skill frameworks for enterprise integrations.

Agentic AI Multi-Agent Context Engineering

Serverless Analytics

Time-series analytics with scalable Lambda workloads and event-driven architecture.

AWS Lambda Serverless Time-Series

How I Work

01

Discovery

1–2 sessions, define goals & acceptance criteria

02

MVP Planning

Minimal production scope with milestones

03

Implementation

Iterative sprints, IaC, automated tests

04

Operations

Observability, runbooks, handover & optional support

Security by Design
Cost Efficiency
Reproducibility
MVP Focus

Skills & Tools

Languages

Python SQL TypeScript Bash

Agentic AI & LLM Systems

Agentic AI Multi-Agent Orchestration Context Engineering Harness Engineering Guardrails Human-in-the-Loop Durable Execution RAG MCP

Frameworks & Providers

LangGraph PydanticAI LangChain LiteLLM FastMCP Temporal Anthropic OpenAI Bedrock

Evaluation & Observability

Langfuse promptfoo Agent Red Teaming Prompt Regression Prometheus Grafana Loki

Cloud & Platform

AWS Azure Databricks Lambda ECS Serverless Terraform AWS CDK Docker GitHub Actions FastAPI

Data Engineering & Lakehouse

Databricks Delta Lake Unity Catalog PySpark MLflow Databricks Asset Bundles DuckDB ETL/ELT Parquet/PyArrow Qdrant PostgreSQL Redis Elasticsearch

Contact

Interested in cloud architecture or MLOps optimization?