Enterprise experience
Internal software, reporting, calculation and decision-support tools built around real business processes in German and international enterprise environments.
Vantura Studios / Portfolio
Software · Data · AI · Product
I builduseful systems.
Software developer and founder of Vantura Studios. I turn real business problems into products, internal tools, data systems and AI workflows.
Discuss a projectProduct ownership
From concept and architecture to implementation, launch, iteration and continued technical ownership.
Business + engineering
Commercial workflows, data, product thinking and hands-on software development in one profile.
Selected work
Proof, not a logo wall.
Public products where possible, transparent NDA cases where necessary, and purpose-built demos where the underlying craft matters more than a client name.
01 / Flagship product
Personal AI workspace & agent orchestration
Grace
A multi-platform AI product designed to coordinate models, specialist agents, tools and workflows in one coherent workspace — with local and cloud execution as first-class concepts.
AI AgentsReact NativeDesktopTool CallingLocal LLMs
Open product
02 / Enterprise / NDA
Reporting, forecasting & calculation tools
Reporting & Decision Tools
Business-critical reporting, forecasting, calculation and decision-support tools developed across German and international organisations. Client data, interfaces and implementation details remain confidential; responsibilities and technical scope can still be discussed.
DataBusiness LogicReportingForecastingInternal Tools
Open product
03 / Client work
Websites, tools & connected workflows
Digital Business Builds
Selected work for smaller businesses and operational teams — from high-trust websites and reporting concepts to connected tools supporting sales, operations and day-to-day decisions.
Next.jsWebReportingAutomationProduct Design
Confidential enterprise work
Screenshots, data and implementation details are intentionally not published. The value of the work can still be described without exposing confidential information.
Real business software does not always belong in a public gallery.
I have developed internal reporting, forecasting, calculation and decision-support tools across German and international organisations. The work combines complex business logic, operational data and iterative stakeholder requirements.
Management reporting & dashboards
Forecasting and variance analysis
Custom calculation logic
KPI and benchmark systems
Data integration & transformation
Internal business applications
Stakeholder-driven iteration
Workflow automation
How I think about the work
Data → Logic → Decision
The interface is only the final layer. The real work is translating processes, rules and imperfect data into something dependable enough to use every day.
01
Sources
02
Transform
03
Business logic
04
Decision
Next proof layer
Live demos.
Synthetic data, real interaction. These demo cases show the type of systems I can build without revealing confidential client material.
Interactive demo concept
Executive Reporting
A management dashboard that turns fragmented operational data into a focused view of performance, forecast, variance and priorities.
Planned showcase
Interactive demo concept
Client Intelligence
A score-card system for account health, growth signals, prioritisation and next-best actions using fully synthetic demonstration data.
Planned showcase
Interactive demo concept
Pricing Engine
A configurable calculation product showing how complex rules, benchmarks, inputs and business logic can become a clean interface.
Planned showcase
AI Agents ✦Agentic Workflows ✦Automation ✦Tool Calling ✦MCP ✦Local LLMs ✦OpenAI ✦Claude ✦n8n ✦RAG ✦Hybrid Routing ✦Human Approval ✦AI Agents ✦Agentic Workflows ✦Automation ✦Tool Calling ✦MCP ✦Local LLMs ✦OpenAI ✦Claude ✦n8n ✦RAG ✦Hybrid Routing ✦Human Approval ✦
AI · Agents · Automation
Agents that work. Automations that stick.
This is the layer I am pushing hardest: agentic systems around real operations — tool-calling agents, local and cloud models, and automations that replace manual handoffs instead of sitting next to them.
How the agent works
Trigger → Agent → Tools → Approval
The chat window is not the product. The product is a controlled loop: something happens, an agent decides, tools execute, and a person still owns the critical step.
01
Trigger
02
Agent
03
Tools
04
Approval
01
AI Agents
Specialist agents that plan, call tools and finish multi-step work with a defined stop.
02
Agentic Workflows
Several agents coordinated around one business process instead of a single chat window.
03
Tool Calling & MCP
APIs, files, browsers and internal systems exposed as tools the agent can actually use.
04
Automation
n8n, scheduled jobs and event-driven flows that replace manual handoffs.
05
Local & Hybrid LLMs
Private data stays local. Frontier models only where they earn the access.
06
Human Approval
Critical actions wait. The system prepares the work; a person still signs it off.
Arsenal
Broad by design.
The useful combination is not one framework. It is moving from a messy process to agents, automations, data and an interface people will actually use.
AI Agents
Agentic Workflows
Tool Calling
MCP
OpenAI
Claude
Local LLMs
n8n
RAG
Python
TypeScript
React Native
Expo
React
Next.js
Supabase
REST APIs
GraphQL
Docker
CI/CD
Power BI
Tableau
Excel
Salesforce
Data Integration
Product Strategy
Project Management
Have a useful problem?
Let's turn it into an agent, an automation, or a system people actually use.
jk@vantura-studios.com
Vantura Studios · AI Agents · Automation · Data