HomeAmericaConstellation Astra Fellowship 2026: Fully Funded In-Person Program Advancing AI Safety Research
Copy of Fellowship Template 1 (1)

Constellation Astra Fellowship 2026: Fully Funded In-Person Program Advancing AI Safety Research

The Constellation Astra Fellowship is a fully funded, 3–6 month, in-person program at Constellation’s Berkeley research center, designed to accelerate AI safety research and talent development. This flagship fellowship aims to bring exceptional individuals into the field of AI safety, connect them with leading mentors, and support high-impact research projects that address urgent and neglected challenges in AI governance, policy, security, and empirical research.

As artificial intelligence advances at unprecedented speed, the Astra Fellowship seeks to prepare researchers and practitioners to tackle some of the most critical risks associated with frontier AI technologies. Over 80% of Astra’s first cohort have transitioned into full-time roles at organizations including Redwood Research, METR, Anthropic, OpenAI, Google DeepMind, and the UK AI Security Institute. The program not only places fellows into high-impact roles but also supports the launch of new initiatives in the AI safety space.


Program Overview

The Astra Fellowship is an in-person program based at Constellation’s Berkeley research center. Fellows work on frontier AI safety projects with guidance from expert mentors and dedicated support from Constellation’s research management and talent team.

Program Duration: 3–6 months
Location: Berkeley, California, USA
Funding: Fully funded, including stipend, research budget, visa support (if required), and workspace access
Application Deadline: September 26, 2025
Fellowship Start Date: January 2, 2026


Who Should Apply

The fellowship is suitable for individuals who:

  • Are motivated to reduce catastrophic risks from advanced AI.

  • Bring technical or domain-specific expertise relevant to AI safety, including technical research, security, governance, policy, strategy, or field-building.

  • Are seeking to transition into a full-time AI safety role or start an AI safety-focused initiative or organization.

  • May have prior AI safety experience, but it is not required. Many impactful fellows have entered from adjacent fields and quickly contributed significantly.

The program encourages applications from motivated individuals even if they do not meet every qualification.


Fellowship Structure

The Astra Fellowship is structured into three main phases:

  1. Orientation & Project Kickoff (Weeks 1–2)

    • Mentors present project ideas.

    • Fellows scope and initiate their research projects.

  2. Research & Career Development (Weeks 2–12)

    • Fellows work closely with expert mentors and Constellation research managers.

    • Receive 1:1 career coaching and recruiting support from the Constellation talent team.

  3. Placement, Incubation, and Extension (Weeks 12+)

    • Fellows publish research outcomes and present findings.

    • Receive support for placement into full-time AI safety roles or incubation of new initiatives.

    • High-performing fellows may qualify for a program extension of up to three months.


Mentorship and Research Areas

Fellows are mentored by top experts from leading AI safety, governance, and policy organizations, including:

Governance & Policy Mentors:

  • Lennart Heim – Independent/AI Policy

  • Michael Chen – METR

  • Additional experts from AI policy think tanks

Security Mentors:

  • Buck Shlegeris – Redwood Research

  • Keri Warr – Anthropic

  • Nicholas Carlini – Anthropic

  • Additional mentors from leading AI security institutes

Empirical Research Mentors:

  • Trenton Bricken – Anthropic

  • Sam Bowman – Anthropic

  • Alec Radford – Anthropic

  • Stephen McAleer – OpenAI

  • Scott Emmons – Google DeepMind

  • Additional senior researchers from Anthropic, OpenAI, Google DeepMind, and UKAISI

Strategy Mentors:

  • Ryan Greenblatt – Redwood Research

  • Julian Stastny – Redwood Research

  • Eli Lifland – AI Futures Project

Field Building Mentors:

  • Alex Fields – Constellation

  • Lauren Mangla – Constellation

Mentorship is delivered through weekly 1:1 sessions, small group meetings, office hours, and Slack collaboration, ensuring close guidance and professional development.


What Fellows Receive

The Astra Fellowship provides comprehensive support to enable full-time research focus:

  • Stipend: Competitive financial support throughout the program

  • Research Budget: Approximately $15,000 per fellow per month for compute resources

  • Workspace & Community: Access to Constellation’s Berkeley research center, with ~150 network participants, daily seminars, workshops, shared meals, and AI safety convenings

  • Mentorship & Research Management: Guidance from senior experts, 1:1 coaching, and career development support

  • Placement Services: Connections to full-time AI safety roles at top organizations

  • Incubation Services: Support for launching new AI safety projects, including business operations, communications, hiring, and fundraising


Success Stories

Previous fellows have achieved significant outcomes, including:

  • Eli Lifland & Romeo Dean: Developed AI 2027 scenarios during Astra, later co-founding the AI Futures Project.

  • Michael Chen: Joined METR’s policy team immediately after the fellowship.

  • Aryan Bhatt: Currently running a research team at Redwood Research following mentorship through Astra.

  • Martin Soto: Works in UKAISI technical staff after extensive learning during Astra.

These stories highlight the fellowship’s ability to launch high-impact careers in AI safety and contribute to transformative research initiatives.


How to Apply

Applications for the Constellation Astra Fellowship 2026 are now open. Applicants are encouraged to apply early as applications are reviewed on a rolling basis.

Application Deadline: September 26, 2025
Fellowship Start Date: January 2, 2026
Apply Here

Applicants should submit proposals demonstrating:

  • Interest and motivation in advancing AI safety research

  • Relevant technical or policy expertise

  • Commitment to making a high-impact contribution to the AI safety field


The Constellation Astra Fellowship represents a unique opportunity for motivated individuals to develop their skills, contribute to critical AI safety research, and transition into impactful roles within the AI safety ecosystem. By combining mentorship, career support, research resources, and community engagement, Astra ensures fellows are well-prepared to tackle some of the most pressing challenges posed by advanced AI technologies.

For more Information and Applications:

Visit HERE

Stay on opportunitiesforwomen.org/ for more opportunities.

Comments are off for this post.

Scroll to Top