Computer Scientist — Adobe Systems Engineering GmbH
Hamburg, Germany | July 2022 – Present (3+ years)
- Developed backbone microservices (Python, Go, Node.js, Java) with a global team; the services handle 100 M+ requests per day across 5 global regions and meet strict low-latency, high-availability SLAs.
- Designed and built agentic AI systems for engineering operations — LLM-powered agents that autonomously investigate production incidents, monitor deployment pipelines, review pull requests, and perform codebase analysis. Integrated 10+ internal tool APIs (Splunk, New Relic, Jenkins, Jira, GitHub, Kayenta, ArgoCD) via custom Model Context Protocol (MCP) servers.
- Authored operational knowledge bases for AI agents — encoded domain-specific troubleshooting workflows, cross-service log correlation patterns, and evidence-based root-cause-analysis playbooks that enable LLM agents to resolve incidents with minimal human intervention.
- Owned CI/CD and GitOps pipelines (Jenkins, Argo Workflows, Helm, ArgoCD) with multi-region progressive delivery; reduced average deployment time from ~120 min to 40 min (-67%).
- Co-maintained an observability stack (Prometheus, Cortex, Grafana, New Relic, Splunk), enabling distributed tracing, near-real-time alerting, and faster incident resolution across all production regions.
- Helped implement automated canary analysis with Argo Rollouts and Kayenta, comparing baseline vs canary via NRQL metric queries, reducing release-related rollbacks and customer impact.
- Acted as Security Champion — tracked CVEs, coordinated timely patches, and kept the stack compliant with internal security baselines.
- Developed ETL pipelines (Python, SQL, PostgreSQL) that consolidate telemetry for critical analytics and ML-driven experimentation.
- Mentored engineers on AI-assisted development practices — introduced LLM-based tooling and agentic workflows to the team, improving developer productivity. Designed and delivered a workshop on RAG pipelines, vector databases, and LangGraph agent orchestration.