Specialist-model research desk

Saved workspace
Direct questions

What is the best next action?

Frontier Development research deskSection 01 · Page 10 of 14Last updated 26 October 2026

Prepare a one-page pilot brief before requesting a production quote. It should state the target task, current system, failed cases, protected holdout owner, acceptance rules, cost boundary, deployment constraint and conditions for stopping.

Use the readiness quiz to locate missing prerequisites, then use the scorecard to compare evidence from suppliers that receive the same brief. A negative feasibility result can be commercially useful: it prevents a larger commitment and tells the team which constraint to repair.

TRACE THE EVIDENCE

Section source library

This page contains no direct external citations. These are the section’s references; they do not independently validate the desk’s proposed scoring or methods.

  1. SFT Trainer — Hugging Face TRL documentation →huggingface.co
  2. Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps — Hugging Face, 3 September 2026 →huggingface.co
  3. A Small Model Asks Better: Fine-Tuning Query Clarification for Retrieval-Augmented Generation — He, Popa and Danilevsky, 19 October 2026 →arxiv.org
  4. Demystifying evals for AI agents — Anthropic, 9 January 2026 →www.anthropic.com
  5. Pricing — Together AI, snapshot checked 26 October 2026 →www.together.ai
  6. Llama 3.3 Community License Agreement — Meta →github.com
  7. A Field Guide to Rapidly Improving AI Products — Hamel Husain →hamel.dev
  8. Guidance on HIPAA & Cloud Computing — HHS Office for Civil Rights →www.hhs.gov
  9. QLoRA: Efficient Finetuning of Quantized LLMs — Dettmers et al., 23 May 2023 →arxiv.org
  10. LoRA Learns Less and Forgets Less — Biderman et al., revision 20 September 2024 →arxiv.org
  11. Extend — Extraction Overview →docs.extend.ai
  12. GroupKFold — scikit-learn documentation →scikit-learn.org
  13. TimeSeriesSplit — scikit-learn documentation →scikit-learn.org
  14. RouteLLM: Learning to Route LLMs with Preference Data — Ong et al., revision 23 February 2025 →arxiv.org
  15. The Open Source AI Definition 1.0 — Open Source Initiative →opensource.org