Christoph Lengowski · AI Product Owner & Senior IT Consultant
Making AI initiatives ready for decisions.
From the first use case to a safe release, I connect product ownership, requirements engineering, and AI evaluation – creating testable requirements, robust prototypes, and clear go/no-go decisions.
Use Cases & AI RequirementsEvaluation & GuardrailsPrototyping & Delivery
8QA team members led
E2EIndependently delivered Playwright test automation
5+Years of delivery ownership
Enterprise AI with Product Focus
From an ambiguous AI idea to an assessable product concept – grounded in business value, clear requirements, and technical feasibility.
AI Evaluation & Quality-by-Design
Defining quality goals, evaluation criteria, and guardrails for AI systems. Risks, failure modes, and probabilistic outcomes become measurable before rollout.
AI Guardrails & Delivery Readiness
Designing traceable evaluation, approval, and feedback loops for AI products. Independently delivered Playwright and TypeScript test automation plus CI/CD experience support reproducible evaluations and safer delivery processes.
AI Requirements Engineering
Translating vague business ideas into prioritized use cases, data and context requirements, suitable RAG, LLM, or agentic architectures, and clear KPIs for probabilistic outcomes.
Rapid AI Prototyping & PoCs
Focused prototypes using local LLMs, RAG, and workflow automation validate value, data access, and technical risk before teams invest in scaling and integration.
AI Product Ownership
Prioritizing AI initiatives by business value, risk, and feasibility. Aligning domain teams, engineering, and stakeholders around measurable outcomes, clear decisions, and realistic delivery steps.
RAG & Knowledge Products
Design and practical evaluation of RAG and knowledge-management products. Personal PoCs using Docker, Open WebUI, Ollama, and LightRAG support informed decisions on retrieval, data quality, and system boundaries.
Certifications
Verified qualifications across product ownership, requirements engineering, AI quality, test management, and structured delivery.
I combine AI consulting, domain analysis, product ownership, requirements engineering, and QA leadership to make robust product decisions.
As a Senior AI Consultant at Sopra Steria, I focus on designing and implementing trustworthy, digitally sovereign AI solutions, particularly for the public sector. I combine AI requirements engineering, AI use-case design, and AI solution design with my experience in IT consulting and quality management.
Since January 2026, I have been building websites and web applications with Codex and Claude Code, increasingly using agentic development at work as well. In February, I founded my startup NutriKompass and developed the entire application myself. This adds hands-on software development and ownership of a product to my consulting experience.
I also maintain a private Personal Context Hub that structures professional experience, projects, decisions, and learning topics as a versioned Markdown knowledge base. It is practical context-engineering work: preparing information for consistent reuse across portfolio work, applications, strategic decisions, and collaboration with AI systems.
Agents & AI Systems
I consider value, system boundaries, evaluation, and operations together: with clear input paths, guardrails, traceability, and verifiable quality criteria.
I translate domain logic, roles, and decision points into usable AI interactions instead of generic chat flows.
I design guardrails, validation, data flows, and traceability into the system from the start.
I build agents as embedded workflow components, not as isolated gimmicks.
I structure and version context as a reusable knowledge base instead of losing information across isolated chats and documents.
Experience
5+ Years Product Delivery
Role
Senior AI Consultant · Sopra Steria
Foundation
M.Sc. + PSPO/IREB/ISTQB
Own product
Founder & developer · NutriKompass
Quick profile
AI product ownership across business, domain teams, and engineering
IREB-based AI requirements engineering for clear use cases and measurable outcomes
Hands-on development of RAG, LLM, and workflow prototypes
Evaluation, guardrails, and quality-by-design as a differentiator
Focus on AI products that remain reliable from PoC to operating reality
What teams value
I bring structure into ambiguous problem spaces and surface risks early.
I treat AI quality as part of product goals, requirements, and architecture – not as a late control step.
I prototype early so product decisions are based on evidence rather than AI promises.
Experience
Consulting as a foundation. Building products as the next step. Since 2026, both paths run in parallel.
Consulting & QA
11/2026 - Present
Senior AI Consultant
Key Impact
Focused on AI requirements engineering, digital sovereignty, and trustworthy AI solution design.
In AI & Generative AI Consulting, with a focus on the public sector, I work on designing and implementing trustworthy, digitally sovereign AI solutions. My goal is to help shape the digital transformation of Germany and Europe and make Generative AI useful for public administration. I combine requirements engineering, IT consulting, and quality management with modern AI technologies.
AI requirements engineering and AI use-case design
Digital sovereignty and trustworthy AI
AI solution design and process optimization
Generative AI consulting for public administration and the public sector
09/2021 - 10/2026
IT Consultant · QA Lead · AI Delivery
Key Impact
Connected business requirements, product ownership, and quality across public-sector IT projects.
From September 2021 to October 2026, I worked at Materna in several roles on IT projects for public-sector clients. My work connected requirements engineering, product ownership, and QA leadership with test management and test automation. In a separate project, as QA Lead and Test Manager, I drove end-to-end test automation with Playwright and TypeScript and implemented it hands-on.
Requirements engineering and alignment between clients, domain teams, and engineering
Product ownership through backlog structuring and prioritization
Functional leadership of a QA team of up to eight people
Test management, risk-based test strategy, and quality assurance
As QA Lead and Test Manager, drove and independently implemented end-to-end test automation with Playwright and TypeScript in another project
Development & founding
Alongside consulting
02/2026 - Present
Founder & developer
Key Impact
My own startup and independently developed application: NutriKompass.
In February 2026, I founded NutriKompass and developed the entire application myself. It brings together product vision, domain requirements and technical implementation.
End-to-end product ownership from idea to web application
Built with Next.js and TypeScript
Agentic development informed by the quality standards of my consulting work
01/2026 - Present
Agentic development & agent building
Since January 2026, I have been building my own websites and web applications with Codex and Claude Code. I also design and build autonomous AI agents with Langdock, and increasingly apply agentic approaches in my day-to-day professional work.
Independently built multiple websites with coding agents
Designed and built autonomous AI agents with Langdock
Brought requirements engineering and QA experience directly into development
Iterative implementation with personal ownership of the result
Working closely with the client, I rebuilt the website from requirements gathering through launch, restructured its content, and optimized it for mobile devices. I also planned, ran, and evaluated a Google Ads search campaign. The project aimed to improve the organization’s visibility among prospective clients and specialist organizations, encourage relevant inquiries, and support long-term occupancy of available places.
How both sides come together
From requirements to application.
Three perspectives, one development process. Explore each step.
Understand
Clarity before the first commit.
I turn a product idea into concrete requirements, user flows and acceptance criteria. This is where my consulting experience comes in.
Selected projects and PoCs showing my path from AI requirements and product concepts to evaluation, guardrails, and technical validation.
01Key Project
Modernization of a Public-Sector Administrative Procedure
Requirements engineering, test management, and independently delivered E2E test automation for a business-critical administrative procedure
AI Product Ownership with Hands-on PoC Practice
Business-oriented use-case prioritization, AI requirements engineering, evaluation, and guardrails – combined with delivery experience and practical RAG, LLM, and workflow validation.
Legacy -> Web
Migration from Oracle Forms/Reports to modern web architecture
MVP
Definition and prioritization of product-oriented scope
E2E
Test automation independently delivered with Playwright and TypeScript
Public Sector
Digital administration with high traceability requirements
Modernization of a business-critical administrative procedure by replacing an Oracle Forms and Oracle Reports landscape with a modern web architecture. In addition to requirements management, process analysis, and MVP shaping, I own test automation end to end – from concept and design to the independent implementation of the automated user journeys.
My role
Test Manager / Requirements Engineer / MVP Co-Lead / Test Automation
Tech Stack
React · Spring Boot · TypeScript · Playwright
Challenge
A business-critical legacy procedure has to move into a modern web architecture without losing domain-specific edge cases, regulatory requirements, traceability, or usability. At the same time, the MVP needs clear boundaries, testable requirements, and a durable quality strategy.
Solution
I connect requirements engineering, MVP shaping, and quality assurance – from user stories, acceptance criteria, and business validation rules to the API-first target architecture. I designed the complete test automation approach and independently implemented it with Playwright and TypeScript. The end-to-end tests are derived from business processes and risks and managed in Jira and Zephyr for traceability to the requirements.
Project context
Modernization of a business-critical administrative reimbursement procedure
Replacement of an Oracle Forms and Oracle Reports landscape with a modern web architecture
Public-sector environment with strong requirements around regulation, traceability, and accessibility
Project scope
Requirements management and derivation of durable requirements from legacy processes and domain logic
Prioritization of user stories, acceptance criteria, MVP scope, validation rules, and edge cases
Complete concept, design, and independent implementation of E2E test automation with Playwright and TypeScript
Derivation of end-to-end tests from business processes, risks, and acceptance criteria
Traceable organization and management of automated tests with Jira and Zephyr
Interface role between business stakeholders, engineering, architecture, and quality assurance
Impact
Analyzed existing business processes and derived durable requirements from legacy systems
Prioritized user stories, acceptance criteria, MVP scope, and domain-specific edge cases
Owned test automation from concept through independent implementation
Designed end-to-end tests with Playwright and TypeScript from business processes, risks, and acceptance criteria
Integrated automated tests into the delivery process with traceability in Jira and Zephyr
Aligned accessibility and regulatory requirements with business, engineering, and architecture
System Architecture
System Flow
Legacy Domain Processes
Requirements & MVP Scope
React / Spring Boot
OpenAPI Interfaces
Playwright E2E Tests
Jira / Zephyr
Core Components
React frontend and Spring Boot backend
OpenAPI-based interface contracts
Playwright and TypeScript test automation with traceability in Jira and Zephyr
Reporting through Jasper Reports
Hard Decisions
Protect domain continuity before replacing technology
Define MVP scope together with testability
Derive end-to-end tests from business risks and acceptance criteria
Guardrails
Anonymize customer-specific details in public material
Test regulatory edge cases and traceability explicitly
Compare legacy and target behavior with risk-based automated end-to-end tests
E-Gov Workflow Platform – Large-Scale QA in the Public Sector
QA Lead / Test Manager for a complex e-government platform
A large-scale e-government platform for digital files and process handling in public administration. Within this complex program, I was responsible for planning, steering, and evolving the entire quality assurance setup, raising QA maturity both operationally and methodologically.
The project ran in a highly complex public-sector environment with multiple clients, backend services, strong traceability requirements, and demanding release expectations. Quality had to be controlled not just through execution, but through risk-based test strategy, defect governance, test steering, and stakeholder reporting.
Solution
I built a structured test organization, led the QA team, and tightly connected test management, automatable test design, and KPI-based reporting with engineering, project leadership, and client stakeholders. I worked with Jira/Xray and Cucumber-based test assets, supported their handoff into Playwright implementation with developers, and embedded the resulting automation in Jenkins- and Bitbucket-supported delivery workflows. In addition, I used AI deliberately through prompt and context engineering to generate, expand, and plausibility-check test cases and test data faster. That turned quality into a controllable delivery capability with clear release readiness instead of a reactive bottleneck shortly before releases.
Highlights
Built a structured test organization inside a large e-government program
Built a risk-based test strategy across multiple system domains
Introduced and expanded Playwright- and Cucumber-based automation in close collaboration with development
Established defect governance and KPI-based quality reporting
Supported integration of automated tests into Jenkins and Bitbucket delivery workflows
Used AI deliberately for test case and test data work through prompting and context engineering
Owned release and regression steering in a complex public-sector environment
Improved release stability through systematic quality steering
Established test KPIs and reporting for leadership and stakeholders
Covered accessibility and data protection requirements as part of the QA scope
Learnings
“Large IT programs do not become stable through good test execution alone, but through strong QA leadership, transparent quality metrics, and tight coordination across team, delivery, and stakeholders. That combination is what I built and owned in this program.”
Making project knowledge easier to find and use in workflows
An internal knowledge system that makes project documents searchable and connects them to existing workflows. Suitable AI use cases were assessed, tested with focused prototypes, and introduced incrementally into QA and requirements processes.
My role
AI Integration / IT Consultant
Tech Stack
LightRAG · Open WebUI · n8n · Lokale LLM-Infrastruktur
Challenge
Enterprise AI initiatives require more than a technically working demo: use cases need verifiable value, sources must be current and dependable, and each PoC needs clear boundaries to support a realistic MVP decision. Without success criteria and process integration, even promising ideas remain difficult to assess.
Solution
I gathered and assessed use cases and defined focused prototypes for suitable topics. This led to a RAG knowledge system with reviewed project documentation. n8n workflows connect Jira, Confluence, and SharePoint to existing processes, while source and retrieval quality were improved with domain teams.
Highlights
Structured AI use case discovery around value, feasibility, data availability, and risk
Shaped PoC and MVP scope with explicit assumptions, boundaries, and review points
Defined measurable success criteria for quality, source coverage, effort, and cycle time
Assessed source quality and curated project documentation into a dependable knowledge base
Built a RAG knowledge system using LightRAG, Open WebUI, and local LLM infrastructure
Integrated n8n, Jira, Confluence, and SharePoint into existing delivery workflows
Learnings
“Enterprise AI becomes investable when use case value, source quality, and success criteria are clarified before the PoC. Technology is only one part; dependable reviews and realistic MVP scope determine whether the result becomes a useful workflow.”