Ammons Data Labs builds pragmatic, production-style systems that bridge applied AI, data engineering, and cloud architecture.
Current focus: PhD Corpus Ingestion & Search — a compact pipeline for extracting, enriching, indexing, and querying a research PDF corpus.
A compact evidence-ingestion pipeline for research PDFs:
- PDF text extraction and cleaning
- Metadata enrichment and normalisation
- Database-backed processing state and repeatable runs
- Full-text indexing for search
- Tests and modular design for extending with APIs, queues, or workers
Reference implementations covering agent observability, .NET services, automation, and domain prototypes (some may not be under active development):
- Observable Agent Starter — observable agent template (DSPy, Langfuse, DeepEval, CI)
- ADL Document Summarization Service — .NET backend patterns for document summarisation
- ADL M365 Automation Starter — invoice automation starter (AI + workflows + CI/CD)
- adl-ba-screening — GIS-backed property screening service (flood overlays, parcel metrics, geocoding)
- HoldThatThread — .NET reasoning API experiment (streaming, clean architecture, tests)
- RoleSkills — skill evidence extraction and scoring from repos and role specs
These repositories are kept for reproducibility and reference and are not under active development.
- Evaluate-SCAIL — encrypted deduplication research codebase (SCAIL/P-SCAIL experiments)