[ THE_INFERENCE_HUB ]
Luma ↗Join the Hub

EVENTS // CALL_FOR_SPEAKERS

Pitch us your talk.

Shipping AI in production? Learned something the hard way? The stage is yours. Apply for a specific event below, or pitch a general topic and we’ll match it to a future one. You’ll track the status of every application in the members portal.

WANTED_TOPICS

Topics the community keeps asking for. Got real experience with one of these? That pitch goes to the top of the pile.

TOPIC_01Agents in production

Real agent deployments: the harness around the model, the failure modes nobody blogs about, and what it took to make them reliable. War stories over demos.

Pitch this topic
TOPIC_02Evals that catch regressions

How you measure LLM quality before your users do. Eval suites, LLM-as-judge pitfalls, and regression gates that actually block a bad release.

Pitch this topic
TOPIC_03RAG beyond the demo

Retrieval quality at scale: chunking, reranking, hybrid search, and what happens when the corpus is messy real-world data instead of a clean benchmark.

Pitch this topic
TOPIC_04The cost of inference

Serving LLMs without burning the budget: caching, batching, quantization, model routing, and how the bill changes as usage grows.

Pitch this topic
TOPIC_05Prompt injection and LLM security

Attacking and defending LLM apps: injection, data exfiltration through tools, and guardrails that hold up outside the slide deck.

Pitch this topic
TOPIC_06Small models, big jobs

When a small fine-tuned model beats an API call: distillation, on-device inference, and picking the tasks where owning the weights pays off.

Pitch this topic

WHAT_WE_LOOK_FOR

Real production lessons over product pitches. Tell us the topic, what the audience walks away with, and the format — a 20–30 minute talk, a lightning talk, a workshop, or a panel. We review every pitch and reply through the portal.