SignalMatch — Explainable AI Engineering
Collection
2 items • Updated
id stringclasses 8
values | lane stringclasses 4
values | job_text stringclasses 8
values | expected_signal_labels listlengths 0 8 | review_note stringclasses 8
values |
|---|---|---|---|---|
ai-platform-full | AI platform | AI Platform Engineer
Build reliable LLM workflows with LangChain, retrieval-augmented generation, tool use, and model evaluation. Own Go services on AWS and Kubernetes. Add regression tests and cost-aware fallbacks. | [
"AI agents and workflows",
"LLMs and generative AI",
"LangChain",
"Retrieval and RAG",
"Evaluation and quality",
"Go and backend services",
"AWS",
"Kubernetes and containers"
] | High-signal AI platform case with eight mapped concepts. |
backend-cloud-full | Backend and cloud | Senior Go Backend Engineer
Design Go microservices and REST or gRPC APIs. Deploy Docker workloads to AWS EKS with Terraform and GitOps. Own observability, retries, and clear SLOs. | [
"Go and backend services",
"Node.js and APIs",
"AWS",
"Kubernetes and containers",
"Infrastructure and GitOps",
"Observability"
] | Cloud delivery case that checks backend, infrastructure, and reliability signals. |
mobile-full | Mobile | Senior iOS and React Native Engineer
Ship reliable product experiences with Swift, SwiftUI, async/await, React Native, Expo, networking, caching, accessibility, and automated release testing. | [
"Swift and iOS",
"React Native"
] | Mobile case with native and cross-platform signals. |
ai-quality | AI platform | AI Quality Engineer
Create evaluation suites for generative AI and LLM applications. Test retrieval, vector search, embeddings, tool-use workflows, regressions, refusal behavior, and latency. | [
"AI agents and workflows",
"LLMs and generative AI",
"Retrieval and RAG",
"Evaluation and quality"
] | Evaluation-focused case for checking quality and safety-oriented language. |
node-platform | Backend and cloud | Node.js Platform Engineer
Build JavaScript backend services with Express and REST APIs. Package them in Docker, deploy to AWS, and improve observability for production workloads. | [
"Node.js and APIs",
"AWS",
"Kubernetes and containers",
"Observability"
] | Node and platform case with container and operations language. |
mobile-cloud-mix | Mobile | Cross-platform mobile engineer
Build a React Native app with a Swift bridge, a Go API, AWS Lambda, Kubernetes deployment, and SLO-based monitoring. | [
"Go and backend services",
"AWS",
"Kubernetes and containers",
"Observability",
"Swift and iOS",
"React Native"
] | Mixed case that checks whether the matcher can surface signals across lanes. |
sparse-description | Manual review | Software engineer
Work with a product team to improve customer tasks, review code, and maintain a growing application. | [] | Sparse case that should remain honest when no mapped technology is visible. |
unrelated-role | Manual review | Senior underwater archaeology historian
Lead field research, preserve maritime artifacts, and publish historical findings with museums and universities. | [] | Intentionally unrelated case that should not produce a false match. |
SignalMatch Evaluation Cases
Small, synthetic job-description cases for testing the public SignalMatch Role Fit Analyzer.
Each JSONL row contains a short role description, the lane it represents, and the concept labels that a deterministic matcher is expected to find. The set includes strong matches, mixed matches, sparse input, and an intentionally unrelated role so that a demo can show both useful coverage and honest uncertainty.
Files
eval_cases.jsonl— eight synthetic evaluation cases.
Schema
id— stable case identifier.lane— review lane used by the demo.job_text— synthetic role title and description.expected_signal_labels— mapped concepts expected in the role text.review_note— why the case is useful for evaluation.
Limitations
This is a small hand-written test set, not a hiring benchmark. It contains no employer names, personal data, application history, resumes, credentials, or private user data. The expected labels describe visible text matches; they do not represent a hiring decision or a claim about a real candidate.
- Downloads last month
- 17