Automation, data, AI
Automation & AI
Services and bots that run without a human in the loop, data processing, RAG and language models — and the Docker and cloud infrastructure they all run on.
What is happening
Pick a scenario — it plays out in the district on the diagram, and every step is explained here.
The animation is slowed down for clarity; latencies in the trace are realistic.
District scenarios
Technologies
Pick a technology to light up the buildings it runs in. Tap or click a building to open its details. Technologies with no building in the city lead to case studies.
In case studies
In case studies
In case studies
- CloudCloud databases and services, connection settings kept outside the code.
In case studies
- RAG & LLMKnowledge-base search and language models: answers grounded in your own sources.
In case studies
- NumPy · SciPy · PandasNumerical methods, interpolation, data processing and Matplotlib plots.
In case studies
- AI-assisted developmentOpenAI Codex and Claude at work: task framing, specifications, review and checking the result.
What it looks like in code
rag/retrieval.py
async def retrieve(question: str, k: int = 4) -> list[Chunk]:
vector = await embed(question)
rows = await db.fetch(
"SELECT id, text, source, embedding <=> $1 AS distance"
" FROM chunks ORDER BY distance LIMIT $2",
vector, k,
)
return [Chunk(**row) for row in rows if row["distance"] < MAX_DISTANCE]