Senior Machine Learning Engineer @ RTL Technology GmbH Köln / Cologne, Germany
I’m a Senior Machine Learning Engineer focused on bridging the gap between promising AI demos and robust, secure internal products. I build the agent platforms, RAG systems, and underlying plumbing that allow non-technical teams to safely leverage AI at scale.
I care about building for business impact and reliability, rather than just chasing the latest models. My day-to-day work sits at the intersection of:
- Agentic Workflows: Moving beyond simple chat interfaces to build multi-agent systems with custom skills, tool-use, and human-in-the-loop approvals.
- Architecture & Performance: Optimizing architectures for local model deployment, managing ultra-low latency services (like speech-to-speech), and ensuring GDPR-aligned data access.
- LLMOps & Governance: Building safe-by-design systems with rigorous prompt-injection defense, tracing, and quality evaluation in production.
- Platform Engineering: Creating self-serve tools that make AI capabilities reusable across the organization.
Previously, I worked as a Data Scientist @ Accenture and Software Engineer @ Tech Mahindra. I hold an M.Sc. in Statistics & Machine Learning from Linköping University (Sweden) and a B.Tech in Electronics & Communication from Rajasthan Technical University (India).
I spend most of my time architecting scalable platforms for internal use. While I can't share the proprietary source code, here is a high-level view of the systems I've brought to production:
A centralized platform enabling people across the company to spin up and deploy specialized agents for daily knowledge work and creative workflows.
- The Tech: Advanced prompt engineering, workflow automation, and custom agent skill development.
- The Impact: Powers specialized agents for data analysis, presentation drafting, and operational assistance for media/video workflows.
A web app builder designed for non-technical stakeholders. It drastically reduces the engineering cycle by allowing users to describe what they need, iterate quickly, and deploy prototype tools without waiting on a traditional sprint cycle.
- Enterprise RAG Platform: Built on LangGraph, Qdrant, and FastAPI, currently powering 100+ internal chatbots with strict access controls.
- Procurement Automation: Developed a multi-agent assistant that contributed to €1M+ in operational savings during its beta phase.
- Production Observability: Implemented comprehensive LLMOps practices (using tools like Langfuse) for tracing, evaluation, and telemetry across multiple active deployments.
- Secure Communications: Architected a multichannel communications generator featuring automated fact-checking and sentiment matching.
I use this profile as a public landing page for the engineering themes I care about most: practical AI agents, GenAI platforms, LLMOps, and tools that help teams move faster.
If you're also working on turning AI into actual products, I’m always open to exchanging ideas, comparing architectural approaches, and discussing real-world adoption challenges.
📫 Best way to reach me: Connect on LinkedIn
I recently curated my public GitHub profile so this landing page better reflects my current focus on GenAI and ML engineering.
Older open projects that still matter—including traditional machine learning, deep learning, time series analysis, and Bayesian modeling—are preserved here: