Job Description

Job Description

🛡️ Help Build the Last Measurable Line of Defense for AI Systems

For 2,500 years, ethical claims have been unfalsifiable.

We're changing that.

Our Research Team is building ErisML (Modeling Language)+DEME (Democratic Ethical Module Engine) — an open-source framework that makes AI alignment measurable, auditable, and actionable . Our core innovation, the Bond Index , provides a single, human-calibrated number that tells you whether an AI system should be deployed.

This isn't theoretical. We've already validated the approach against 7 historical disasters — from the $504M Citibank-Revlon incident to the Mars Climate Orbiter loss to the Uber ATG fatality. Every one would have been detected.

The window for building robust AI verification tools is closing fast

The systems we're trying to verify are getting more capable every month.

We need volunteers who believe alignment must be engineering, not philosophy — and who want to help build the tools before we need them.

What We're Building:

• ErisML Core — Formal language for agents, norms, and multi-agent interaction
• DEME — Democratically Governed Ethics Modules with 9 ethical dimensions
• Bond Index — Quantitative coherence verification with witness generation
• MCP Integration — Works with any MCP-compatible AI agent
• HPC Integration — Integration, and Foundation model research and testing using the resources of SJSU CoE HPC Lab:

What You'll Do:

Depending on your background, you might:

• Extend categorical mathematics foundations (functors, natural transformations, cohomology)
• Build integration pipelines for real-world ML training systems
• Identify adversarial cases and deployment failure modes
• Develop domain-specific transform suites (healthcare, finance, autonomous systems)
• Ground ethical invariances through democratic deliberation frameworks
• Improve documentation, write tutorials, or translate materials
• Review code, report bugs, and help others get started

This Role is For You If:

✅ You believe AI safety requires measurable verification, not just principles

✅ You want to contribute to open-source infrastructure that matters

✅ You can commit 2-5+ hours/week (flexible, async-friendly)

✅ You're comfortable with ambiguity and building something new

Commitment: Flexible, volunteer basis. Most contributors work 2-10 hours/week asynchronously.

Location: Remote / Global (Note: — The HPC can only be accessed locally, on the campus network)

Compensation: This is an unpaid open-source volunteer position. Contributors are credited in publications and releases.

Requirements

We're looking for contributors across multiple tracks:

Track A: Technical Contributors

• ML Engineers, Software Engineers, or Research Engineers
• Familiarity with Python, PyTorch/JAX, or formal verification tools
• Interest in AI safety, alignment, or robust ML

Track B: Mathematical/Theoretical Contributors

• Background in category theory, type theory, topology, or formal methods
• Interest in extending theoretical foundations
• PhD students, postdocs, or researchers welcome

Track C: Safety & Policy Contributors

• AI safety researchers, red-teamers, or policy analysts
• Experience identifying failure modes or adversarial scenarios
• Interest in governance frameworks for AI systems

Track D: Domain Experts

• Practitioners in healthcare, finance, autonomous vehicles, content moderation, or other high-stakes domains
• Can provide real-world use cases and validation scenarios
• Help ensure the framework addresses actual deployment challenges

Track E: Ethics & Deliberation Contributors

• Ethicists, political theorists, or deliberative democracy researchers
• Interest in grounding machine ethics in democratic processes
• Help develop the G_declared invariance specifications

Track F: Community Contributors

• Technical writers, documentation specialists, educators
• Anyone who can help make this accessible to more people
• No deep technical background required — enthusiasm and clarity matter

Minimum for all tracks:

• Genuine interest in AI safety and alignment
• Ability to collaborate asynchronously (GitHub, Discord)
• Willingness to learn and ask questions

How to Apply

1. ⭐ Star the repo: github.com/ahb-sjsu/erisml-lib
2. 📖 Read the DISCUSSIONS_WELCOME.md
3. 💬 Introduce yourself in GitHub Discussions
4. 🔧 Pick up a good-first-issue or propose your own contribution

Questions? Reach out to andrew.bond@sjsu.edu

"The Bond Index is the deliverable. Everything else is infrastructure."

Ethical Finite Machines — Ordo ex Chāōnā; Ethos ex Māchinā

#AIAlignment #AISafety #OpenSource #MachineLearning #AIEthics #ResponsibleAI #Python #GitHub #TrustworthyAI

Job Tags

Similar Jobs