Headline

AI Engineer | Voice Bots · Agent Assist · RAG Systems | Python · LangGraph · FastAPI | From PoC to Production

About

I spent the last two and a half years at Sogedes GmbH building the backend that kept their conversational AI products running in production. Voice bots, agent assist, chatbots — the kind of systems where a bad response or a slow retrieval is immediately someone's problem. What I actually worked on: - Agentic workflows: Multi-agent orchestration and state-machine routing from scratch using LangGraph and Python, handling mid-conversation logic that had to stay predictable under real traffic - Backend services: Containerised FastAPI microservices deployed on Azure and GCP, designed to stay stable across concurrent workloads - Retrieval systems: Session-based RAG using ChromaDB and Redis, tuned for low latency and fewer hallucinations in live conversations - Data privacy: Automated PII extraction and masking before data ever reached the model, using Microsoft Presidio and Hugging Face NER I have a Master's in AI and Data Science from Technische Hochschule Deggendorf. The degree gave me the theory. Sogedes gave me the habit of shipping things that actually work. I like working on problems that are hard enough to be interesting, in environments where I can own something properly from the first design decision through to production. Based in Mannheim. Open to roles across Germany and remote.

Location

Mannheim

AI stack

antigravity

claude code

claude

codex

cursor

deepseek

Work experience

Management Trainee AI/ Conversation AI at Kapture CX, May 2026 - Present · 4 mo, Bengaluru, Work on voicebot and chatbot implementations, configuring and enhancing conversational workflows based on business requirements • Develop and optimize voicebot call flows, including re-attempt logic, automated callbacks, and conversation-flow enhancements • Configure and enhance chatbot conversation flows, responses, and interaction logic • Perform functional and scenario-based testing of voicebots and chatbots to validate conversational flows and expected behavior • Analyze voice recordings and call conversations to identify response, flow, and interaction issues and recommend improvements • Perform debugging, troubleshooting, validation, and regression testing for bot enhancements • Support deployment and rollout of conversational AI enhancements across client implementations

Education

Master of Science - MS, Artificial Intelligence and Data Science, Deggendorf Institute of Technology

Master of Science - MS, Artificial Intelligence and Data Science, University of South Bohemia in České Budějovice