AI Agents, Systems, and Automation
I develop AI agent systems and explore their application in real-world workflows. My focus is on building tools and applications that solve practical problems — from automated research pipelines to multi-agent development frameworks.
Areas of Work
AI Agents
Specialized agents for research, analysis, content creation, code generation, and data processing
Multi-Agent Systems
Coordinated agent workflows with handoffs, quality checks, and human oversight at critical checkpoints
Automation Pipelines
End-to-end automated pipelines — from data collection through processing to output generation, running on schedule or on demand
Tool and API Integration
Connecting agents to external systems — GitHub, Jira, databases, APIs, and file systems
Open to select collaborations on small, well-defined projects.
Independent Projects
Systems and applications built with the AI Development Framework — each one exploring a different aspect of agent-based development.
AI Development Framework
ActiveOpen-source framework orchestrating multiple AI agents for automated software development. Ticket-based workflows with Implementation, Test, and Review agents collaborating with human oversight.
CardCroc
ActiveCross-platform flashcard app for iOS and Android with Spaced Repetition (SM-2), Firebase Auth, multilingual support, and a vocabulary pipeline producing 20 learning decks with ~10,000 cards.
TradingOctopus
ActiveAI-driven financial data platform with FastAPI, TimescaleDB for time-series data, multi-source data mining (Yahoo Finance, Finnhub), and automated market data collection.
ResearchRaven
ActiveAI research-to-content pipeline in three stages: Research, Knowledge Organization, and Content Generation. Grounded in communication theory with parallel multi-agent research.
TicketFalcon
Early StageChat-based AI assistant for writing structured tickets. Connects to knowledge sources (Confluence, Git) and uses LLMs to formulate well-structured tickets directly in Jira.
fdrechsler.de
ActiveThis website — built entirely using the AI Development Framework. Next.js, Docker, CI/CD, comprehensive analytics, and full German legal compliance.