Research Internship 2026

Part time

Remote

Voluntarily

Applications for the Research Internship are now open. The deadline for applications is 23:59 EOD Sunday 21st June 2026.

Job Description:

About the Internship
The HUX AI Research Internship is a volunteer-based, project-driven program designed for students and early-career professionals who want to explore real-world applications of responsible AI and data governance.

Interns will join HUX AI for a 10-week period to work on hands-on, impact-oriented micro-projects aimed at increasing AI literacy, operationalising governance frameworks, and enabling safer AI adoption across sectors.

Interns will receive mentorship and feedback from the HUX AI team, engage in weekly check-ins, and contribute to high-quality, practical deliverables that support public and organisational understanding of emerging AI standards.

Research Internship Program Start: 6th of July 2026

Project Tracks (Candidates can express a preference or be assigned based on fit):

Project 1 — LLM Guardrails Evaluation Toolkit
Purpose: To develop a practical toolkit for evaluating whether LLM guardrails can resist prompt injection, jailbreak attempts, policy bypass, unsafe tool-use triggers, and related failure modes.

Scope: The project will focus on defensive evaluation of LLM guardrails. It may include a risk taxonomy, safe test prompt library, scoring rubric, control matrix, and lightweight evaluation template.

Expected outputs:
- Guardrail failure taxonomy
- Safe test prompt library
- Guardrail effectiveness scoring rubric
- Risk-control-evidence matrix
- Lightweight evaluation tool or template
- 5–10 page report
Project 2 — Agentic AI Risk Control Matrix

Purpose: To create a practical governance framework for AI systems that can use tools, access data, trigger workflows, or perform actions.

Scope: The project will focus on autonomy levels, tool-use risks, human oversight, access control, auditability, action reversibility, and required governance controls.

Expected outputs:
- Agentic AI risk taxonomy
- Autonomy level classification
- Scenario library
- Agentic AI control matrix
- Risk assessment questionnaire or tool
- 5–10 page report

Eligibility and Selection Criteria

We welcome applications from undergraduate and graduate students, recent graduates, and early-career professionals from a wide range of disciplines, including but not limited to: engineers, computer science, data science, law, design, business, public policy, and international relations.

We are especially looking for individuals who demonstrate:

Experience in developing an open-source project
Strong organisational or research skills
Curiosity about LLMs, Agentic AI, responsible AI, governance, and standards
Clear and effective writing
Ability to work independently and meet deadlines
Openness to feedback and collaboration

Prior experience in AI or policy is required, but an eagerness to learn and contribute is essential.

Duration and Commitment

Internships are part-time and remote, offered in a 10-week track.

All interns will:

Join a welcome orientation
Attend weekly project check-ins with mentors
Receive feedback from HUX AI mentors (primary and guest mentors)
Be acknowledged in the final project outputs

This is a volunteer-based internship. However, select outstanding contributors may be invited to collaborate on future paid projects or to receive formal recommendation letters from HUX AI leadership.

Application Process
Step 1: Submit your application form
Step 2: Shortlisted candidates will be invited to a 10/15 minute video call
Step 3: Final selection and onboarding

Application Deadline: by EOD on the 24th of June 2026.

You can review the 2025 projects and reports here.

You can watch short interviews about the 2025 projects here.

Follow us on Instagram.