Jev and System One Models in Practice

**Why Attend**

Most AI systems today run on System Two: slow, expensive, generative LLM calls that reason at length. But most decisions inside engineering pipelines don’t need that. Is this test result suspicious? Which code issue matters most? Should this workflow act or escalate?

Those are System One decisions: fast, repeatable and structured, and your software needs to act on them immediately.

Jev is a System One decision model. You provide state, ask typed questions, and get back structured decisions with confidence scores that your application can branch, rank or route on directly. It is built to work alongside System Two LLMs, not replace them.

This workshop is about System One models in practice. **Instead of an API demo, you’ll build three projects **across testing, code health and observability. You’ll see when a System One decision is the right tool, when to hand off to an LLM, and how to combine the two in production.

**Meet Your Instructor**

**Renata Andrade**

**Quality Engineer Lead | Test Automation Engineer & Architect**

**Renata **is passionate about software quality and she helps teams build the best strategy to practice quality in the easiest, fastest and most reliable way possible, adding more value to the business. She is driven by excellence, innovation and creativity and loves making a difference, learning, sharing knowledge, and inspiring people.

###

**Three Projects. Three Real Uses of Jev.**

Project 1: AI-Powered Exploratory Testing with Jev + Playwright

Connect Jev to a live application and Playwright workflow to make fast, structured judgments about application behaviour.

You’ll learn how to turn observations from an exploratory test into decisions your automation can use immediately, including confidence-aware verdicts and escalation.

Project 2: Code Health Triage at Scale

Engineering teams can generate far more static-analysis and code-health findings than developers have time to investigate.

Using output from tools such as RepoWise / CodeScene, you’ll use Jev to classify, score, and prioritise findings so attention goes to the issues that matter most.

Project 3: AI Observability: Measure What’s Actually Working

Adding AI to a pipeline does not automatically make the pipeline better.

In the third project, you’ll examine token cost, model usage, agent effectiveness, decision quality, and orchestration overhead to understand whether an AI-powered workflow is delivering value.

You’ll learn how Jev can participate in online evaluation and operational decision-making rather than simply generating more AI activity.

**What You’ll Take Away**

You’ll learn how to

• Explain the difference between System One and System Two models, and when each fits
• Use Jev’s typed decision primitives for classification, scoring and confidence-aware decisions
• Design hybrid pipelines that use System One decisions by default and escalate to LLMs only when needed
• Turn AI decisions directly into branching, ranking and routing logic
• Set confidence thresholds that decide whether software acts or escalates
• Measure the cost, speed and reliability gains of moving decisions from System Two to System One

**Who Should Attend**

• **Software Developers** building AI-powered workflows and applications
• **QA and Test Engineers** exploring AI-assisted testing
• **DevOps and SRE Engineers** working on triage, routing, and automation
• **AI Engineers** designing hybrid Jev + LLM systems
• **Engineering Leaders** looking for measurable AI efficiency rather than AI experimentation

**What you’ll receive**

• Two-hour live, instructor-led workshop
• Three working projects covering testing, code health, and AI observability
• A reusable repo containing the workshop code
• Complete workshop recording
• Workshop slide deck
• Live Q&A with the instructor
• Certificate of completion

###