Direct engagements · Remote-first · B2B

Agentic AI in production
no proof of concept.

I'm Sascha Willomitzer, working in enterprise environments for 24 years — from J2EE through CTO responsibility to agentic AI systems. I build and operate agentic AI systems, cloud platforms and Kubernetes infrastructure for companies that need results in production — not slide decks about results.

20,000 agentic calls per day — the system I built for an energy provider and handed over to production
2 months from zero to live — complete replacement of the legacy solution, built from scratch
24+ years of enterprise experience — from J2EE through real-time advertising to agentic AI
25+ mandates since 2018 through my company — before that, CTO, architect and developer

Four focus areas. One profile.

My work sits at the intersection of AI engineering, platform architecture and operations — exactly where agentic systems have to become production-ready.

Core

Agentic AI & LLM Engineering

Agentic telephony and chat systems, RAG pipelines with hybrid retrieval, tool calling, guardrails, evaluations and LLM orchestration. Production operation including monitoring, incident handling and cost/latency optimisation — with numbers that are measurable in production.

Foundation

Cloud & Platform Architecture

Azure landing zones (hub/spoke), on-premises-to-cloud migrations, OpenShift → AKS, PostgreSQL high availability with under 10 minutes of downtime. Zero-downtime design for production-critical systems.

Operations

Kubernetes & DevOps

On-premises Kubernetes in high-security environments (air-gapped vSphere), Cilium as CNI including Gateway API, Terraform, ArgoCD, GitOps. Operations, not just build-out — including knowledge transfer to your teams.

Industry focus

Regulated Industries

Medical devices (ISO 13485, MDR, FDA), critical infrastructure (KRITIS, BSI), public-sector clients (with passed security clearance). I know the specific requirements that public-sector and regulated clients place on traceability and documentation.

Production proof

Four projects that show what my work delivers — from agentic AI systems to a government platform. All numbers from real mandates, all systems in production.

05.2026 – today

Agentic AI platform for an energy provider

Senior Manager Software AI-Engineering (interim)

In-house development of an agentic conversational-AI layer behind BabelForce (telephony), connected to Zendesk (ticketing/CRM). Complete redevelopment and replacement of the legacy solution within two months, then scaled from ~1,000 to ~20,000 agentically handled calls per day. OpenAI and Anthropic APIs, tool calling, RAG, STT/TTS, guardrails, monitoring — plus an eight-person team I built that now runs the system independently.

  • Kotlin / Spring AI, LLM orchestration, low-latency design
  • DevOps ownership: OpenTofu, ArgoCD, GitLab CI
  • AI-assisted development process: agentic code reviews as standard
09.2024 – 06.2025

Cloud transformation in the medical device sector

Senior Cloud Consultant

Azure landing zones in a hub/spoke architecture, migration of on-premises OpenShift to AKS, PostgreSQL migration via pgBackRest and Crunchy PGO to CNPG — with effectively under 10 minutes of downtime. Regulated environment: ISO 13485, MDR, FDA, validation requirements.

  • Zero-downtime design for production-critical systems
  • IaC with Terraform, GitOps with ArgoCD
  • AI-driven development with Claude Code
12.2025 – today

Learning platform in a high-security environment

DevOps & Application Engineer

Building, training and operating a learning platform for a public-sector client — Kubernetes (Tanzu) in an air-gapped vSphere private cloud, CNI network policies, network segmentation, Kyverno, Teleport. Project environment with passed security clearance.

  • Air-gapped operation, disjoint security domains
  • CI/CD, GitLab, containerisation in high security
10.2014 – 01.2016

CTO — AI in credit risk management

Chief Technology Officer, ~40 staff

As CTO of a financial-data provider, responsible for probabilistic search models (fuzzy search) in credit risk management: daily index updates from the Oracle database, processing fraud-detection, positive, negative and neutral data. Reorganised the company, 150 % more development throughput, daily instead of monthly deployments.

  • Machine learning in a financial context — in production since 2014

Stage experience

Architecture has to be explainable — to teams, to management, to 10,000 people in an arena. My talk “Knights & Rockets — Focus, Value, Innovation” shows how to communicate technical topics honestly and concretely.

“Knights & Rockets”

From a weather-computing system for spaceflight to burnout as an IT director to founding my own company: a talk about why focus, real customer understanding and the courage to experiment are what decide outcomes — not tools.

Watch on YouTube →

Direct — no detours

You book me personally: a senior expert with production ownership, no hidden junior team. My company carries the mandates, the back-office is taken care of — your projects get my undivided attention across architecture, implementation and operations.

Email

info@thechameleonway.com

Reply within one business day.

Phone

+49 (0) 2133 / 868 123-0

Personal line, no call queue.

Contract & company

Engagement offers via The Chameleon Way IT GmbH →

B2B direct engagements, ISO-certified, EU hosting.