München, 10.05.2026

Vladislav Vorobev

Foto VAVPC
80469 München
☎ +49 1772597575
✉ j@vavpc.de

Developer / DevOps Engineer — AI | LLM | Agents | Infrastructure | Kubernetes | Hardware

Delivers across the stack — backend, infra, hardware, ML/LLM workloads — with AI agents and AI pipelines as everyday tooling for codegen, refactors, test scaffolding, and code review; modern, agent-driven development as the default workflow. Foundation in Python, Rust, TypeScript, distributed systems, and Kubernetes, including bare-metal, hardware-controlled clusters. Ships production LLM agents (chat, RAG, tool-using, MCP) and LLM/RAG evaluation pipelines. Modern concepts like: code costs nothing, no human review.

Competencies and Skills (Selected)

  • LLMs, Agents & Eval Ops: OpenAI, Claude, Hugging Face — tool/function calling, structured outputs (Pydantic, JSON Schema), streaming, vision, prompt caching; LangChain, LangGraph, PydanticAI, CrewAI, OpenHands, plain-Python tool-using agents (ReAct, planner-executor, sub-agents); MCP servers and clients; eval pipelines (golden datasets, hallucination detection, prompt regression, LLM-as-judge, eval in CI); guardrails, jailbreak resistance; cost/latency budgets, observability (traces, token spend).

  • RAG & Retrieval: FAISS, pgvector, Weaviate; sentence-transformers / OpenAI embeddings; hybrid search (BM25 + dense), reranking, chunking, citation grounding.

  • Agent-Driven Dev: full automation of the dev process — multi-agent workflows with specialised roles (planner, developer, tester, code reviewer, security reviewer, ops/deploy); orchestrator + sub-agents, parallel sub-agents, planner-executor, reviewer/critic patterns; agents read specs, write code, open PRs, run evals, review and merge; humans set direction only.

  • Languages & Frameworks: Python, TypeScript / JavaScript, Java, Groovy; FastAPI + Pydantic, FastStream, asyncio; SvelteKit / Svelte, React — native stack; prior production in Django, Pyramid, Flask, Tornado, Vert.x, Grails, Rails. Any other language or framework on demand via AI agents.

  • Architecture: satellite (stable core + specialised satellites for infra, ML, hardware control, automation, integrations); microservices, event-driven, CQRS, data mesh.

  • APIs & Protocols: REST, gRPC, WebSocket, SSE; OpenAPI, AsyncAPI, Protobuf; SSH, SSL/TLS; MQTT (IoT).

  • Data, ETL, Big Data: databases — PostgreSQL + PostGIS, ClickHouse, ScyllaDB, Redis, MongoDB, MySQL/MariaDB; messaging & streaming — Kafka, Redis (Streams + pub/sub); processing — Spark/pySpark, Delta Lake, Airflow; data libs — Pandas, Arrow; object storage — MinIO.

  • Cloud, DevOps & CI/CD: Kubernetes (incl. bare-metal), Docker, Terraform, Nomad; AWS, Azure, GCP, Hetzner; CI/CD (GitHub, GitLab, Jenkins, Gitea).

  • Kubernetes: kubeadm bare-metal clusters, air-gapped; RBAC; CRDs and operators with custom controllers; CNI (Calico, Cilium / eBPF); MetalLB, Ingress / Egress, cert-manager; Rook-Ceph (PVC, PV, StorageClass) + CSI snapshots; Deployments, StatefulSets, Jobs, CronJobs; autoscaling (HPA, VPA, cluster autoscaler); Prometheus operator (ServiceMonitor, PrometheusRule); descheduler, nvidia-device-plugin; cluster lifecycle, upgrades.

  • Observability: Prometheus + Alertmanager, Grafana, Loki + Promtail, Tempo, Grafana Alloy (OpenTelemetry collector); auto-metrics via node-exporter, kube-state-metrics, kubernetes-event-exporter, metrics-server.

  • Networking & Hardware: bare-metal servers (HPE, Dell), OOB management (iLO, iDRAC, IPMI, Redfish); enterprise switching (Cisco) on leaf-spine fabrics; firewalls (Palo Alto); load balancers — hardware and Nginx, HAProxy, Traefik (software); WireGuard / VPN; iptables / nftables, NAT; datacenter ops — rack-and-stack, structured cabling (copper / fiber), optics (SFP+, QSFP), patch panels; firmware / OS patching and lifecycle.

  • Methodology & Practices: AI-driven validation — agents apply and enforce TDD, BDD, DDD, schema validation, Clean Code, SOLID; spec-driven development; API-first / contract-first design; 12-factor app; GitOps.

  • Quality: unit, integration, end-to-end, contract, property-based tests; CodeQL, SonarQube, Sentry.

  • Tools: Neovim, VS Code, PyCharm, Cursor, Zed, gh, kubectl, helm, k9s, docker, terraform, aws CLI, gcloud, hcloud, psql, redis-cli, etc.

  • Domain: Automotive, Telematics, IoT, Sensors, Battery, Audit & Compliance, E-Commerce, Mapping / Geo, Properties.

Non-IT

  • Business & Stakeholders: stakeholder management; business context; requirements & task scoping; technical-to-business translation; trade-off communication.

  • Team Leadership: distributed / async teams; mixed human + AI-agent workflows; outcomes-based delivery; international teams (incl. remote, 09/2025 – present). Prior: Scrum, SAFe.

  • Communication: async written-first (RFCs, specs); cross-functional alignment; mentoring on AI workflows.

Professional Experience (Selected)

Google – DevOps Engineer, Google Distributed Cloud Hosted (GDCH)

08/2025 – present, Munich, consulting

  • DevOps for Google Distributed Cloud Hosted — air-gapped, on-premise Kubernetes for sovereign and regulated workloads.

  • Hardware controlled by Kubernetes: bare-metal bring-up (HPE, Dell, iLO / iDRAC / IPMI / Redfish), fleet lifecycle; network gear integration (Cisco switching, Palo Alto firewalls).

  • Custom controllers, CRDs, and operators (Go) for hardware operations and lifecycle automation.

  • Cluster upgrades, observability; agent-driven ops via runbooks and IaC-as-truth.

    Tech: Kubernetes (CRDs, operators, custom controllers), Go, Linux, Bare Metal (HPE, Dell), iLO / iDRAC / IPMI / Redfish, Cisco, Palo Alto, GCP.

DesignWeltDeko – AI Engineer & Full‑Stack Developer, E-Commerce

12/2024 – 08/2025, Munich, consulting

  • Shipped customer-facing and internal AI agents — Product Chat (RAG over catalogue), Scene Generation, Text Generation, Image Classification — with LLM/RAG evaluation (golden sets, hallucination checks, prompt regression) in production.

  • Connected a custom ERP (no existing connector) to Shopify via event-driven sync; products, prices, AI-classified images flow in real time. Integrated eBay, Etsy, Amazon.

  • Shipped a Progressive Web App (React, Svelte) and Liquid storefront customisations.

  • Operated bare-metal + Hetzner Kubernetes (Calico, MetalLB, Ingress-Nginx, Rook-Ceph) hosting platform, vector store, ERP; observability (Prometheus, Grafana, Loki, Tempo, OpenTelemetry) and ETL (Kafka, Airflow).

  • Coached the team on LLM agents, prompt engineering, and AI-assisted workflows.

    Job-specific tech: Shopify (REST/GraphQL), Liquid; eBay / Etsy / Amazon APIs; Google Analytics / Merchant / Ads / IDP; Hetzner ↔ on-premise.

SAP SE – Software Engineer, Automated Audit Framework

06/2024 – 12/2024, Walldorf (consulting)

  • Delivered new compliance features in the audit framework; upgraded core dependencies (Python, Pyramid).

  • Refactored legacy modules — reduced LoC, raised performance.

  • Continuation of earlier SAP self-audit platform work (see 2019–2022 entry); focus on maintainability, forward compatibility, Data Mesh patterns.

  • OAuth / SSO integration; collaborated with multiple SAP business-unit teams.

    Tech: Python, Pyramid, FastAPI, REST, OAuth / SSO, PostgreSQL, MinIO, Docker, GCP, Kubernetes, GitHub Actions.

Mercedes‑Benz AG – Senior Python Developer

01/2023 – 05/2024, Stuttgart (permanent)

  • Optimized micro-service applications for battery test data analysis and executed cloud migration.

  • Migrated battery-test analytics to Azure Blob and Databricks; reduced processing time.

  • Improved unit-test coverage.

    Tech: Python, FastAPI, Pydantic, Pandas, Arrow, Delta Lake, Kubernetes, Azure (MIC/eXtollo internal clouds), Azure DevOps.

SAP SE – Lead Dev, Distributed Self‑Audit Platform

05/2019 – 12/2022, Walldorf (consulting)

  • Architected micro‑services ecosystem collecting audit data from 30+ SAP business units.

  • Implemented MinIO‑based storage ingestion handling 12 TB/day in production.

  • Designed PostgreSQL schema & tuning; backup.

    Tech: Python, asyncio, Kafka, MinIO, Kubernetes, AWS/GCP.

Skoobe GmbH – Big Data Engineer (Python/Java)

05/2018 – 12/2021, Munich (consulting)

  • Ported monolith services from Python 2.7 and Java to async Python micro-services; reduced latency.

  • Built GraphQL gateway consolidating 3 legacy APIs; cut maintenance effort.

  • Supported migration to AWS, automated CI/CD with Jenkins and Terraform.

    Tech: Python, Java, FastAPI, GraphQL, AWS, Jenkins, Nomad, Terraform.

MAN Truck & Bus AG – Backend Developer (Data Processing)

01/2018 – 05/2020, Munich (consulting)

  • Processed terabytes of CAN‑Bus data in AWS S3 using pySpark; enabled real‑time safety alerts.

  • Automated the data classification pipeline.

    Tech: Python, Spark, S3, Parquet, AWS Glue, Airflow.

Mozaiq Operations GmbH – IoT Dev & DevOps Engineer

01/2016 – 12/2018, Munich (consulting)

  • Developed a backend platform for IoT interconnectivity.

  • Built cloud connectors for Bosch/Siemens and Osram etc; expanded device coverage.

  • Deployed Vert.x micro‑services on GCP Kubernetes in production.

    Tech: Groovy, Java, Vert.x, GCP, Kubernetes, MongoDB.

Earlier Consulting & Leadership (2009 – 2015)

Team Lead, Head of IT roles, E‑Commerce platform builds, GIS startup; details on request.

Personal AI-Driven Projects

pois.world — Open map & virtual-parcel platform (2025 – ongoing) — https://pois.world

Solo, fully AI-agent-driven build: six microservices, seven shared libs, ~71k LOC.

  • Concept → Plan → Spec → Code → Feature discipline; spec / feature registry the agents read from and write back to — knowledge synthesised across the codebase, not stuck in any one head.

  • pois-mcp MCP server exposes the platform to AI assistants.

  • FastAPI + SQLAlchemy async + PostGIS (~1M parcels); SvelteKit + Svelte + MapLibre GL JS / CesiumJS; ScyllaDB chat; Kubernetes in production.

Education

Dipl.-Ing. Elektrotechnik (Informationstechnik) – TH-Köln, 2009

Lomonosov Moscow State University – Faculty of Computational Mathematics and Cybernetics, 2006

Languages

German (native) | English (fluent) | Russian (native)

Note

Detailed project portfolio with longer project descriptions and client references available on request.