Skip to content
Open to projects & collaboration

Jack Tigges

Full-Stack & AI Engineer

Started with Minecraft plugins, arrived at enterprise software and AI agents. Never stopped building in between.

Scroll
About

The person behind the code

Jack Tigges
Jack TiggesMunich, Germany
Based in
Munich, Germany
Born
2005
Years coding
5+

I'm 20 and based in Munich. It started in 2016 with Minecraft plugins, because a game should do something it could not do on its own — on the side at first, and seriously and regularly for a good five years now. That question turned into a career: through Python and server administration, through client work on the web, all the way to SAP development in an enterprise context.

Today I build on two fronts. Enterprise by day: ABAP, Fiori, ERP processes. Software that runs at the core of a company and forgives no mistakes. Alongside that, Myrodex, my own startup, where we build a management platform for esports organisations, including our own AI agent with tool access to the platform data.

What I like about the work is the moment a system stops being complicated, when the right abstraction is found and the rest follows by itself. That is why I care about architecture, about tests, and about someone else still understanding the code a year from now.

Away from the screen I train a lot: competitive endurance sport and regular strength work at the gym. Both for the same reason as programming. It is about working patiently on something that only becomes visible after months.

Stack

What I work with

Tools are replaceable, the thinking behind them is not. These are the ones I reach for.

Languages

08
  • TypeScript
  • JavaScript
  • Python
  • Swift
  • Dart
  • ABAP
  • Java
  • SQL

Frontend

08
  • React
  • Next.js
  • Tailwind CSS
  • Flutter
  • SwiftUI
  • Three.js
  • HTML
  • CSS

Backend & data

09
  • Node.js
  • Bun
  • PostgreSQL
  • Supabase
  • Drizzle ORM
  • GraphQL
  • REST & WebSockets
  • Auth & RBAC
  • CloudKit

AI & agents

06
  • Agent-Orchestrierung
  • Claude & OpenAI API
  • LangChain
  • LLM-Observability
  • RAG & Embeddings
  • MCP

Enterprise & SAP

04
  • SAP ERP
  • SAP Fiori
  • ERP-Prozesse
  • Anforderungsanalyse

Infra & tooling

10
  • Docker
  • Linux
  • Git
  • CI/CD
  • CLI & DX-Tooling
  • npm-Publishing
  • Vercel
  • Playwright
  • Sentry & OTel
  • Claude Code
CoreStrongWorking
Experience

Where I have been

The stations, and what I built at each.

  1. 2016 until 2017Project

    The first line of code

    Minecraft plugins · At home

    Started coding at eleven, because the game should do something it could not do on its own.

    • Server plugins in Java: custom commands, items and game mechanics
    • Configuration and persistence in files instead of by accident
    • The first real users were the other players on the server
    • Java
  2. 2017 until 2018Project

    From the game to the system

    Python & server administration · Personal projects

    Moved to Python and my own servers. The first code that was not tied to a single game.

    • Built tools, bots and automation in Python
    • Set up my own Linux servers and kept them running
    • Backups, monitoring and networking basics, usually after the first outage
    • Python
    • Linux
  3. 2018 until 2022Freelance

    First clients, first deadlines

    Client projects & web development · Freelance

    Websites and smaller applications for paying clients, alongside school.

    • Web presences for Quinosens GmbH and the vehicle rental company Quinomobil
    • Responsive frontends, deployment, domains and certificates
    • First backends of my own: SQL databases, APIs, authentication
    years of client work
    5
    • JavaScript
    • HTML
    • CSS
    • SQL
    • Node.js
    • Linux
  4. 2023 until presentProject

    Custom agents and automation

    AI engineering & agents · Personal projects

    Building with language models since 2023, not just writing with them.

    • Agents with tool access, multi-round tool calling, structured output
    • RegRadar DACH: a multi-agent system that watches EU regulation and writes briefings
    • Jarvis: a local-first macOS assistant, cloud only for the hard cases
    EU frameworks tracked
    4
    years building agents
    3
    • Python
    • TypeScript
    • Agent-Orchestrierung
    • Claude & OpenAI API
    • LangChain
  5. 2023 until 2026Employment

    SAP & enterprise development

    Apprenticeship: IT specialist for application development · Munich

    Apprenticeship focused on long-grown ERP systems, where mistakes are expensive.

    • ABAP for business logic, SAP Fiori for the interfaces
    • Custom transactions and reports for the business departments
    • From gathering requirements through to production release
    final grade
    1,4
    years in ERP
    3
    • ABAP
    • SAP Fiori
    • SAP ERP
    • SQL
  6. 2026 until presentFreelance

    Co-founder & Developer

    Myrodex Richter, Schreit & Tigges GbR · Munich

    My own startup: a platform that runs an entire esports organisation.

    • Teams, matches, leagues, sponsors, inventory and helpdesk in one multi-tenant system
    • Myro: an AI assistant that proposes actions instead of executing them on its own
    • Playwright E2E, Sentry and Langfuse tracing from day one
    modules in one platform
    9
    founders
    3
    • TypeScript
    • Next.js
    • React
    • Bun
    • PostgreSQL
    • Drizzle ORM
    • +3
Live data

AI tokens burned

I build with AI agents instead of talking about them. This counter reads my real Claude Code sessions. No estimates.

Live · Total tokens

10,759,081,880

≈ 42,761,366 words written · roughly 107,591 books' worth

Sessions
2,897
Messages
64,527
Projects
45
Input
720,329
Output
57,015,154
Cache written
284,407,650
Cache read
10,416,938,747

By model

  • Sonnet 56B55.8%
  • Opus 53.5B32.6%
  • Fable 5763M7.1%
  • Fable 5.1314M2.9%
  • Opus 5.5172M1.6%

Activity over the last 30 days

02 Sept01 Oct

Commit activity

since 2023

Commits
1,217
Repositories
43
Active days
159
Longest streak
51

Languages by share of code

  • TypeScript31.4%
  • JavaScript20.5%
  • Java19.4%
  • Python10.7%
  • Swift6.3%
  • Other11.8%
How this worksUpdated this minute

A Node script reads the JSONL logs of my local Claude Code sessions, sums the token fields per model and per day, and writes an anonymised snapshot into the repo. The site loads it through an API route. Project names and conversation content never leave my machine.

Contact

Let's build something

A concrete project or a loose idea, either way just write.

Send an email

Status: Open to projects & collaboration