Stage

Inference

Serving models locally. NIM, Triton, TensorRT-LLM — what each replaces in your API bill, what each costs in cold-start and memory.

Article №13 observability NeMo Evaluator ~60 minutes end-to-end — 40 s to ingest the blog into pgvector, 2 min for retrieval, 4 min for generation across three 8B variants, 90 s for the LoRA variant, 9 min for grading

Ragas, Reranked — What 44 Held-Out Questions Say About the Second Brain Stack

A Ragas-style harness written in 200 lines of stdlib Python, run locally on the DGX Spark, against four variants of the Second Brain RAG chain. Naive RAG scores 3.30 / 5. Rerank RAG scores 4.27. LoRA+RAG is a surprise — it does not beat naive. Retrieval is where the points come from.

Article №09 inference NeMo Guardrails ~90 minutes on top of the article #7/#8 chain

One Rail, Three Policies — NeMo Guardrails on the Retrieval Path

NeMo Guardrails drops a policy gate between retrieval and generation. One install, three per-arc configs — PII for Second Brain, style for LLM Wiki, code-safety for Autoresearch — and a 15-query benchmark: 100% block recall, 100% clean pass. Rails are scaffolding; detectors are the content.

Article №08 inference Llama 3.3 70B + Nemotron-Super-49B + Llama 3.1 8B NIM ~30 minutes on top of the article #7 chain

Bigger Generator, Same Grounding — 8B vs 49B vs 70B on One Retrieval Chain

Article #7 bet that a bigger generator would heal the 8B Google-IPO refusal. Ran the A/B across three sizes on one retrieval chain. Bet lost: Nemotron-Super-49B over-refuses the 8B baseline; Llama 3.3 70B narrows the gap, not closes it. The refusal was the scaffold working.

Article №07 inference Nemotron Reranker + pgvector full-text + Llama 3.1 8B NIM ~45 minutes on top of the article #6 chain

Hybrid Retrieval on the Spark — BM25, Dense, Fusion, Rerank

Four retrieval modes on one corpus — naive dense, BM25, Reciprocal Rank Fusion, Nemotron rerank. Dense is already 92% recall@5; rerank adds a point at K=10 and reorders the top. The 8B generator still refuses where retrieval is perfect — grounding, not retrieval, is the new bottleneck.

Article №06 inference Llama 3.1 8B NIM + Nemotron Retriever + pgvector ~30 minutes if the three endpoints are already warm

Three Endpoints, One Answer — Naive RAG on a DGX Spark

Three endpoints in one curl chain — a query embeds through Nemotron, pgvector returns top-5 chunks in under 80 ms, and a Llama 3.1 8B NIM stuffs them into a strict-context prompt. The chain works; the 8B generator still refuses on questions its own context answers.

Article №05 inference pgvector ~15 minutes first install, re-runs in seconds

Where Your Vectors Live — pgvector on a DGX Spark

The substrate between the embed call and the retrieve call — pgvector 0.8.2 running as a Postgres 16 container on GB10, with 1000 Nemotron vectors, HNSW and ivfflat both indexed, and a planner that prefers seq scan until you tell it otherwise.

Article №04 inference NeMo ~30 minutes first install, ~1 minute every restart after

Your Own Semantic Space — a Nemotron Embedding NIM on a DGX Spark

The embedding endpoint that every downstream RAG, wiki, and agent piece will reuse — a 2048-dim Nemotron Retriever NIM running locally on GB10, ready 52 seconds after docker run and holding 28 docs/s under batched load.

Article №03 inference NIM ~2 hours first install, ~2 minutes every restart after

Your First NIM on a DGX Spark — What 24.8 Tokens Per Second Doesn't Tell You

First-contact notes on NVIDIA's DGX-Spark-specific Llama 3.1 8B NIM. 9.4 GB image, ~108 s warm-cache cold-start, 24.8 tok/s steady, OpenAI-compatible on :8000 — and a confidently wrong Python one-liner that clarifies what small-model FP8 buys and what it costs.