Keep tool results inside a token budget
Bound every model-visible tool result before it enters the next prompt.
Dotnet
Your First Qdrant Collection
Run Qdrant locally, connect with the official .NET client, and create a collection whose vector size and distance match the embedding model.
Run production evaluations outside the user response path
Hand production samples to a separate evaluation worker instead of making users wait for an evaluator model.
Attach build identity to operational telemetry
Stamp telemetry with the immutable build that produced it so incidents can be compared across releases and rolling deployments.
AI Systems Need Runtime Budgets
How to bound time, tokens, retries, tool calls, and estimated cost across a complete AI execution instead of limiting each call in isolation.
Use FrozenDictionary for stable, read-heavy lookup data
Build stable lookup tables once with FrozenDictionary, reuse them across reads, and measure whether lookup savings justify the construction cost.
Return Retry-After when a client should back off
Tell API clients when another attempt may succeed instead of leaving them to guess a retry delay.
Version prompts, models, datasets, and evaluation results together
Keep every evaluation result tied to the exact prompt, model identities, dataset, evaluator, and configuration that produced it.
Choose histogram boundaries from the service objective
Put the latency objective into the histogram boundaries so compliance can be calculated without interpolating between buckets.
Provider Independence from Day One
A C# implementation for keeping provider coupling at explicit boundaries while making model capabilities and provider-specific behavior visible.