courseMachine Learning

Jev by TypeSafe AI: System One Models & Open-Source Alternatives

Curated by @tripp

Updated Sep 23, 2026

Learn Jev, TypeSafe AI's System One model, plus its open-source alternatives Laya and Von. This hands-on course covers the /v1/systemone API, the Choice, Score and Noul questions, confidence routing, TypeSafe's cookbooks, calibration (ECE), and CEO Diogo Almeida's "Prod, not God" thesis. You finish by building a triage pipeline that runs on either the hosted or a self-hosted backend.

#jev#typesafe#system-one-models#laya#von#calibration#llm-routing#open-source-ai

Models used

Models the creator uses in this course — chat, study guides, flashcards, ingest. Updated when the listing is republished.

    Syllabus

    Folder tree from your library — expand modules to see lessons.

    View full syllabus

    Changelog

    • v1.02026-09-24

      The course is complete. Module 05 adds a decision guide comparing Jev, Laya, Von and LLMs with structured outputs, with a cost/latency calculator and hybrid architectures, and a 60-minute capstone: a swappable ticket-triage pipeline with a mock evaluation harness. This release also fixes the syllabus practice prompt and the Laya server citation.

    • v0.32026-09-23

      Added Module 04, on open-source alternatives. The Laya lesson covers its architecture, how to read its calibration tables, its benchmarks against Jev, its first-party laya-serve /v1/systemone server, and a local demo. The Von lesson covers the protocol clone and includes a contract checker. The syllabus now lists recommended prerequisites from the public AI Engineering and RAG courses.

    • v0.22026-09-23

      Added Module 03: "Prod, not God". It covers CEO Diogo Almeida's Latent Space episode, with a chapter map, RLCD vs RLHF/RLVR, the refusals debate, traction claims, and a critical read of his claims against the launch materials. The syllabus prerequisites are now self-contained.

    • v0.12026-09-23

      First release: syllabus, Foundations (System One models and calibration) and Module 02, Jev in practice (API, the three question types, confidence routing, cookbooks and known limits). The CEO, open-source and capstone modules are coming next.

    Attribution

    LearnOS is a learning workspace, not a content host. This listing curates external resources; original creators retain ownership. Learners may need to obtain materials separately where noted.

    Command Palette

    Search for a command to run...