Reading paths

Choose the problem you are working on. Each path explains where to start and connects practical tools with the experiments behind them.

• Document Processing and Dataset Quality: Follow a document from collection through extraction, preparation and validation.
• Retrieval, Embeddings and Evidence Tracking: Check representations, source versions and what a retrieved result actually supports.
• AI Agent Systems: Memory and Evaluation: Follow state, attribution and evaluation through small agent systems.
• LLM Interpretability and Evaluation: Move from a model view to compatible captures, interventions and bounded comparisons.
• Neural Architecture Experiments: Separate implemented mechanisms, small pilot findings and untested architecture proposals.