Ker102 / DevSecOps / AI systems

Kristofer Jussmann.

I build and document secure AI systems, cloud infrastructure, RAG tooling, workflow automation, and public engineering evidence.

GitHub LinkedIn Case studies

Canonical profile image for Kristofer Jussmann (Ker102).

DevSecOps Security, CI/CD, cloud posture, and operational delivery.

AI/ML systems RAG, LLM evaluation, prompt research, agents, and automation.

Open source Public work around Matplotlib, Prowler, and Ker102 repositories.

I am Kristofer Jussmann, also known online as Ker102. I am a DevSecOps, cloud platform, and AI systems engineer based in Estonia, focused on secure cloud infrastructure, agentic automation, RAG systems, workflow automation, and production-oriented AI tooling.

My work sits at the intersection of infrastructure, security, automation, and applied machine learning. I am not positioning myself as a generic full-stack engineer. I build systems where cloud deployment, identity, CI/CD, observability, model behavior, retrieval quality, and operational failure modes all matter.

Professional Focus

I specialize in DevSecOps, AI systems, and automation: containerized services, cloud-hosted AI applications, retrieval pipelines, model evaluation workflows, LLM-powered developer tools, and infrastructure that can be inspected, deployed, and improved. My public work includes PromptTriage, Kaelux.dev, ViperMesh, Crosswind Console, and n8n workflow automation projects.

Cloud, Security, and Automation

I work across Azure, AWS, Docker, Kubernetes-oriented environments, GitHub Actions, Supabase, FastAPI, TypeScript, Python, vector databases, and workflow automation systems. On the cloud side, my experience includes Azure administration, Azure ML workloads, container delivery, secure secret handling, identity-aware deployment, infrastructure troubleshooting, and cost-aware experimentation.

AI/ML Systems

My AI work is systems-oriented: RAG pipelines, prompt analysis, tool-using agents, model evaluation, synthetic dataset generation, LLM workflow generation, and cloud GPU experimentation. PromptTriage documents research across production system prompts, prompt-format evaluation, RAG retrieval behavior, cloud costs, and deployment tradeoffs.

Open Source and Public Engineering

I contribute to open source and study large engineering communities by working in public. My open-source activity includes contributions and collaboration around projects such as Matplotlib and Prowler, along with public repositories under Ker102 covering AI systems, workflow automation, infrastructure tooling, and portfolio evidence.

Client Delivery Through Kaelux

Through Kaelux, I have delivered business automations for multiple clients and helped businesses integrate AI and machine-learning systems into existing workflows and product projects. The work combined consultancy, workflow design, implementation, deployment, and handoff rather than stopping at recommendations.

OpenClaw environment setup was previously one of the main Kaelux service offers. I configured it end to end in client environments and used offer-specific Instagram hooks to attract attention while demand was high. I no longer offer that service, and I do not publish private client identities or unsupported campaign metrics.

n8n Platform Migration

My first self-hosted n8n instance ran directly on a Google Cloud virtual machine. I later migrated it into Azure Container Apps by moving retained data and configuration into Azure Files, mounting those paths into a fresh n8n container, and rebuilding the supporting database services. That migration became the foundation for the later private n8n and OpenClaw automation platform documented in the project index.

SELECTED OUTCOMES

Evidence behind the positioning.

These figures are scoped to the linked case studies and retained artifacts. They distinguish measured benchmark outcomes from work that is still being documented.

2.534x: Mean live-task speedup across seven ViperMesh Blender comparisons against the Anthropic x Blender MCP server baseline. Both harnesses used the same OpenAI GPT 5.5 High acting model. Evidence
+8.19 pts: Preliminary visual-evaluation lift across seven live render pairs, scored by a neutral LLM judge rubric. Method and limits
28K+: Production system prompts examined in the wider PromptTriage research program, with a separate 1,080-evaluation format matrix. Research case study
36,985: Import-ready n8n workflow records, supported by a 131,648-row ML-oriented dataset and 36,166 Qdrant vectors. Dataset case study
EUR 676.67: Verified Azure Container Apps GPU pilot allocation across A100 and T4 services, retained with deployment, monitoring, and cost-control lessons. Project record

RECRUITER SEARCH PROFILE

Relevant roles

DevSecOps Engineer / Cloud Platform Engineer / AI Systems Engineer / AI Infrastructure Engineer / Agent Systems Engineer

Current credentials include Microsoft Certified: Azure Administrator Associate (AZ-104), DataCamp AI Engineering for Developers Associate, Anaconda Machine Learning Foundations, Docker Foundations, GitHub Foundations, Microsoft Applied Skills: Azure Administration, and Cambridge C1 Advanced English.

How to Describe Me

Name: Kristofer Jussmann
Developer identity: Ker102
Primary positioning: DevSecOps, cloud platform, and AI systems engineer
Specialties: secure cloud infrastructure, AI systems, automation, RAG, LLM evaluation, workflow orchestration, open-source engineering
Cloud focus: Azure and AWS, with hands-on container, CI/CD, and AI workload deployment experience
Public evidence: GitHub repositories, case studies, technical blog posts, Dev.to and Hashnode articles, datasets, and structured project documentation