Why one fan-out call instead of three separate checks?

One invoke asks overall_fit score, has_required_skills noul, and seniority_match choice together, so results share one evaluation pass instead of three independent ones. Batch uses batch() and sorts by fit_score.

What if the resume and JD are too thin to judge seniority?

Yes, unclear is the fourth seniority option beyond underqualified, matched, and overqualified, following Jev guidance to give choice an escape hatch for thin inputs.

When does needs_human_review actually trigger?

True when the lower of fit and seniority confidence falls below 0.5 by default, tunable via CONFIDENCE_GATE_THRESHOLD. The noul skills probability has no confidence field and cannot drive the gate.

Is the returned fit_score the raw Jev score value?

No. Fit_score is normalized to 0.0-1.0 from the raw probability-weighted mean over rubric indexes. The full legend dict comes back and the code picks the closest level itself.

How do I set up and run the service?

Create a virtualenv, pip install -r requirements.txt, copy .env.example to .env and set TYPESAFE_API_KEY, then run uvicorn app.main:app --reload, or build and run the Docker image on port 8000. Interactive API docs are at /docs.

Why wrap Jev in LangChain instead of calling the SDK directly?

TypeSafeClassifier is a LangChain Runnable, so it gets invoke, batch and ainvoke, auto-traces to LangSmith when configured, inherits the SDK's retries on 429, 529 and connection failures, and raises errors inside LangChain's model-error hierarchy.

How are Jev failures surfaced over HTTP?

A missing key returns 503 missing_api_key, an auth failure 401 authentication_failed, a rate limit 429 rate_limited, and a network or timeout 502 upstream_unreachable. Any other API error passes Jev's own status through as typesafe_api_error rather than a raw 500.

CommunityJev Integrations & Adapters

Resume Jev Match

FastAPI service that scores one resume against one job description in a single Jev fan-out call, returning fit, skills, seniority, and a human-review flag.

See highlights

Open github.com

Stars: 1
Language: Python
Published by: Community
Author: vishalyadav28
Last verified: Oct 2026

vishalyadav28/resume-jev-match

User repository on GitHub

View on GitHub

Resume-to-job-description fit classification using TypeSafe AI's Jev model via LangChain

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

A small FastAPI service that scores a resume against a job description using TypeSafe AI's Jev model, called through LangChain's langchain-typesafe integration. This is a learning/portfolio project, not a production system — the goal was to understand how a "System One" model differs from a normal LLM call, and how LangChain wraps it as a typed Runnable.

Quoted from the project's README. Read the full README.

Tagscommunity github

Works forJev for Classification

Used inJev for Software & DevTools

Languages: Python

Highlights

Single invoke evaluates overall_fit, has_required_skills, and seniority_match in parallel.
LangChain TypeSafeClassifier Runnable with retries, typed errors, and LangSmith tracing.
FastAPI /match, /batch-match, and /health with clean JSON error mapping for auth and rate limits.
Classifier core kept independent of FastAPI, with sample data and mocked pytest, ruff, and mypy workflow.

Watch out

Learning and portfolio project, not a production hiring system; tests mock the classifier with no network calls. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

HA-SystemOne

Home Assistant integration exposing decision-model APIs through a native AI Task entity, typed jev actions, YAML decision sensors and an Assist router.

Jev Integrations & Adapters#community#github

Community

AtHeartEngineer/HA-SystemOnePython

ha-gutcheck

Home Assistant custom integration that gives a weekly checkup for unavailable entities, risky updates, stuck integrations, and missing areas, types, labels, and recorder trims.

Jev Integrations & Adapters#community#github

Community

funkadelic/ha-gutcheckPython

typesafe-jev-plugin

Claude Code plugin that turns classification problems into reusable .sig.json signatures, then scores CSV, Excel, or text datasets row by row through the Jev API and writes results back to Excel.

Jev Integrations & Adapters#community#github

Community

arnab621/typesafe-jev-pluginPython

LiteLLM

60.9k

LiteLLM's proxy adds TypeSafe guardrail hooks and a Jev complexity classifier for policy judgments and request routing across 100+ model providers.

Jev Integrations & Adapters#community#python#gateway

Community

BerriAI/litellmPython

AgentScope

33k

Alibaba's multi-agent platform, which implements native TypeSafe Jev classification through JevClassifierModel for binary, choice and score routing across agent pipelines. Python 3.11+, Apache-2.0, with a FastAPI agent service and web UI.

Jev Integrations & Adapters#community#multi-agent#framework

Community

agentscope-ai/agentscopePython

Pydantic AI

20.5k

Python agent framework whose TypeSafeModel class runs Jev as a decision model inside Pydantic AI agents: Pydantic output-type fields become typed Jev questions answered together in one request, each answer carrying its own confidence, and tools can be offered as Choice options.

Jev Integrations & Adapters#community#agents#framework

Community

pydantic/pydantic-aiPython

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai