À propos de ce poste AgenTrust Fellow (Inaugural Cohort) chez Opaque Systems
About OPAQUE
OPAQUE is the Confidential AI company. Born from UC Berkeley’s RISELab, we solve the core challenge blocking AI adoption at scale: security concerns about data leaks or compliance violations. OPAQUE provides verifiable privacy and governance for AI so organizations can safely run models, agents, and workflows on their most sensitive data. Its Confidential AI platform delivers verifiable runtime governance backed by cryptographic proof that data, models, and agent actions remain private, governed, and compliant with approved policies throughout every AI workflow. This extends traditional data governance tools with real runtime verification, enabling teams to responsibly deploy AI using their most valuable proprietary data, and move from pilot to production 4-5X faster. Customers and partners include ServiceNow, Anthropic, Encore Capital, Accenture, and leaders across high tech, financial services, insurance, and healthcare.
Learn More at Opaque.co
Read about our Values at Opaque.co/about
ABOUT THIS FELLOWSHIP
The AgenTrust Fellowship is a six-month, paid, part-time program for engineers, researchers, and governance practitioners to help build the open standards for verifiable AI agents. Fellows work alongside OPAQUE Chief Platform Officer and Agent Governance Toolkit creator Imran Siddique, contributing to the open agentrust-io ecosystem: TRACE, Confidential MCP, Confidential A2A, Agent Manifest, Model Manifest (Weight Custody), and AGT.
It is a community and standards program, not an internship, research grant, or hiring pipeline. Fellows keep their current role or studies and contribute about 20 hours a week.
Logistics:
6 months (September 15, 2026, to March 2027). Part-time, about 20 hrs/week. Fully remote. Stipend $30,000 to $50,000 per fellow for the term. Cohort of 3 to 5. Applications open August 1 and close August 29, 2026, via Greenhouse.
What fellows do
- Contribute real work to one home track: open standards for agent identity, protecting models and data during execution, and cryptographic evidence of how an agent behaved.
- Ship at least one merged PR or spec contribution, plus one public technical artifact (blog, spec doc, or conference talk).
- Meet weekly 1:1 with Imran and monthly as a cohort. Rishabh Poddar (CTO) is available on request for attestation and TEE questions.
Who should apply
- Security engineers who have red-teamed agentic systems and want to build the containment layer, not just document the problem.
- PhD students and postdocs in formal verification, distributed systems, or attestation.
- AI governance and compliance practitioners who can turn what auditors ask for into machine-executable policy.
How to apply
- Please fill up the form below and
- Submit a 2-page proposal grounded in a specific repository within our ecosystem. Your proposal should reference a specific open issue or spec gap you've identified
- Include your GitHub handle so we can review your history and contributions. Note: public contributions (issues, PRs, discussions) count toward your application, so prior engagement with the repo works in your favor.
- Indicate which home tracks you are interested in.
- Share relevant public work and/or contributions.
