The year 2025 marks a turning point in industrial software development: The bridge between classic engineering and the world of Large Language Models (LLMs) has been built. My focus during this period was not only on technical implementation but also on strategic consulting for companies during their AI transformation.
Consulting & Strategic Advice:
- Use Case Analysis: Identification of value-adding AI application scenarios in existing business processes, from automated documentation to intelligent code analysis.
- Model Selection & Deployment Strategies: Advice on choosing between cloud-based models (OpenAI/Anthropic) and local open-source alternatives via Ollama, with special consideration of data protection and IP protection.
- Architecture Review: Evaluation of existing software architectures for their “AI-Readiness” and support for the integration of LLM interfaces.
AI-Supported Development & Engineering:
- GitHub Copilot: Deep optimization of development workflows in VS Code. Use of Copilot not only as an autocomplete tool but as a sparring partner for complex refactoring and Test-Driven Development.
- Ollama: Local deployment and management of models like Llama 3 or Mistral to realize high-performance AI features without external dependencies.
Frameworks & Autonomous Agents:
- LangChain (Python & JavaScript): Orchestration of complex Chain-of-Thought workflows. Development of AI agents that can independently solve tasks within defined boundaries.
- Retrieval-Augmented Generation (RAG): Conception and construction of RAG systems that combine corporate knowledge (vector databases) with the eloquence of LLMs. This enables precise answers based on internal, non-public data flows.
This experience allows me to support companies holistically: From the first strategic idea to the selection of the appropriate architecture to the productive implementation of secure and efficient AI systems.