Machine learning

Turn a decision the user can't answer alone into a Markdown questionnaire for a specific person (PM / designer / coworker) to fill in async.

Build a persistent compounding wiki from raw sources using the LLM Wiki pattern

Use for debugging DSPy programs, inspect_history, tracing LLM calls, custom callbacks, observability, monitoring, and cost tracking.

Use for DSPy retrieval with dspy.Embedder, dspy.Embeddings, FAISS indexes, semantic search, and local or hosted embedding models.

Use for GEPA reflective optimization, ReAct agent optimization, feedback metrics, LLM reflection, and execution trajectories.

Use for RAG pipelines, retrieval augmented generation, ColBERTv2, context retrieval, multi-hop RAG, and grounded DSPy answers.

Reserved and on-demand GPU cloud instances for ML training and inference.

Test vector stores for embedding inversion, cross-tenant leakage, and poisoning.

Deploy Llama Guard, NeMo Guardrails, and LLM Guard input/output scanners as runtime defenses.

Identify poisoned training data and backdoored models across the ML pipeline.

De-identify clinical text for drafts, notes, and manuscript excerpts using phi-cleaner before sharing with external tools or collaborators.

Evaluate Anvia agents and retrieval — deterministic metrics, semantic similarity, LLM judges, RAG quality, and CLI eval runs.

Use when working on Cosmos world model serving, inference, serverless training smoke validation, backend selection, or rendering limitations.

Use when recommending how to run a workbench workload, writing user-facing docs, or answering "what should I run first?" — lead with the Nebius GPU path and…

Use when choosing or reviewing GPU targets for NPA workbench tools, training, rendering, inference, or workflow YAML resources.

Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and task-specific model training

Perform finance web searches and local context searches.

Analyze finance text sentiment using FinBERT or LLM.

Design tenant-scoped memory, retrieval and retention for agents.

Optimize content for AI search and LLM citations across AI Overviews, ChatGPT, Perplexity, Claude, Gemini, and similar systems.

Vovk.ts JSON Lines streaming — generator handlers (`function*`, `async function*`), the `iteration` validation option, `validateEachIteration`, the…

Building LLM tools with Vovk.ts — `deriveTools()` (procedures → tools) returning a `StandardToolV0[]` array, standalone tools via the `standard-tool` package…

Build retrieval and RAG with Anvia — chunking, embeddings, vector stores, knowledge graphs, filters, and search tools.

Check whether the stated problem is the right problem before work commits to it.