Research Fellow

Role Overview

We are looking for researchers who can do original work and carry it from an open question to a finished result. This role is for someone with at least 3 years of excellent, hands-on research experience. We are hiring across the range of SAFL's research, and we expect to hire more than one person.

SAFL's research runs across two broad areas, which are set out below. You do not need to work across all of them. We want people who are deep in one and can speak to the others.

What every fellow shares is the ability to take a question nobody has answered, choose a method that fits it, do the work, and write up a result that others can check. You will work closely with the research leads and have room to shape your own work and the wider agenda.

This is our most senior research bar. For people with significant relevant experience, we can also hire for a senior version of this role. We also encourage exceptional people who do not have this experience to apply.

SAFL is in a field that moves quickly. The work is demanding, and we mean that plainly. You will take open questions that nobody has answered yet and carry them to a result, often with no clear precedent to follow. Priorities will change as the field changes, and you will need to change with them. Your work will be read by funders, government bodies, and researchers outside the lab, so the standard is high. If you want a role with a fixed scope and a settled routine, this is not that role. If you want to test yourself on hard problems and grow quickly, you will find that here.

About SAFL

The Secure AI Futures Lab (SAFL) is South Asia's first research and capacity-building lab for navigating powerful AI. SAFL pursues parallel interventions in AI governance and impacts research, expert talent advancement, government engagement, and ecosystem support infrastructure. Our foundational work received support from Schmidt Sciences and the Future of Life Institute, among others.

Our Research

Our research follows the same two areas we publish under.

Managing AI Risks

Study the challenges that come from advanced AI, and work out how to reduce them.

• Technical AI governance and policy. The governance frameworks advanced AI needs, and the technical artifacts those frameworks depend on. This covers evaluation standards, incident reporting, logging and provenance, attribution methods, liability and accountability, and the work of turning technical evidence into standards that Indian and international bodies can adopt.
• AI security and alignment research. The safety and security of AI systems themselves, including AI and cybersecurity, and the security of the systems that AI is deployed on top of, such as India's digital public infrastructure.

Societal and Economic Impacts

Understand how AI spreads through society and the economy, and what it changes.

• Economic impacts of AI. How advanced AI reshapes labour, growth, and public finances, with attention to India and other developing economies.
• Psycho-social impacts of AI. How AI changes the way people live, work, and relate to institutions and to each other.

How we work across both

Evaluation, red-teaming, and measurement are the methods we lean on most, whichever area you sit in. Our current projects show this. We are mapping the evaluations, red-teaming infrastructure, and attribution methods that agentic AI governance in India requires, and we are building India-specific benchmarks to evaluate Indian language models. You do not need to arrive with these methods, but you should be willing to learn them.

Key Responsibilities

Research

• Take an open question in one of SAFL's areas and carry it through to a result that others can check.
• Choose methods that fit the question, whether that means an evaluation, a dataset, an economic model, a legal or policy analysis, or a literature synthesis.
• Be rigorous about evidence. Say what your work shows, and say what it does not.

Writing and Output

• Write up methods and results as research notes, reports, or papers for both technical and policy readers.
• Help turn findings into material that can inform Indian AI governance bodies and other partners.

Collaboration

• Work with academic partners, frontier labs, AI safety institutes, and SAFL's wider network.
• Support the team in shaping the broader research agenda.

Internal Operations and Documentation

• Keep clear internal records of research code, data, and results.
• Contribute to reusable methods and documentation.

Who We're Looking For

The ideal candidate may not have all of these. If you are on the fence, please apply.

Experience and Track Record

• At least 3 years of excellent, hands-on research experience, shown through work that others can check, for example published research, a policy report, an open-source evaluation suite, or an empirical analysis.
• A strong foundation in a field relevant to one of our areas. This includes computer science, machine learning, security, economics, public policy, law, or a relevant social science. It can come from a strong degree program or from an equivalent self-taught record of real work. We do not require a PhD.
• A record of taking a research question from an open state to a finished, defensible result.

Skills across all areas

• Clear writing. You can explain a method and a result so that others can follow and check it.
• Careful reasoning about evidence, including what a piece of evidence can and cannot support.
• Quantitative thinking. You reason with numbers, and you check them.
• The ability to read across both technical and policy material and find the main point.
• Comfort using AI tools such as Claude to speed up your own work, while always checking the output for accuracy.

Skills depending on your area

• For technical AI governance, and for AI security and alignment research, strong Python and the ability to write clean, reproducible code for experiments. Experience with large language models and with agents that use tools. Comfort with evaluation and red-teaming methods, and with frameworks such as Inspect.
• For policy work within technical AI governance, close familiarity with how AI governance is actually made, and the ability to write for readers inside government.
• For economic impacts, strong applied empirical methods, and comfort with data work and with modelling.
• For psycho-social impacts, strong empirical social science methods, whether quantitative, qualitative, or both.

Working Style

• High agency. You are comfortable shaping loosely defined problems and deciding what to build.
• Careful and rigorous about methods and evidence.
• Comfortable with ambiguity and with revising your work, since this is a young lab in a fast-moving field.

Good to Have

None of these are required but they will make you stand out.

• Published research or open-source work in AI safety, security, interpretability, evaluations, AI governance, or the economics of AI.
• Experience building dangerous-capability evaluations of AI agents, or red-teaming model APIs.
• Experience contributing to open evaluation frameworks such as Inspect.
• A background in machine learning systems or performance engineering.
• Experience writing for policymakers, or work that has been used by a government body or regulator.
• Familiarity with the work of global AI safety institutes, and with AI safety fellowships and courses.
• Familiarity with India's AI ecosystem and with its shared digital systems, often called digital public infrastructure or DPI, such as UPI and Aadhaar.

Internship Track

This role has an internship track for applicants who do not yet have the research experience above, including current students and recent graduates. The research work is the same and the term is shorter. If that is you, apply through the same form and mark that you would like to be considered for the internship track. It is a way to join the lab and grow into the full role. Interns are paid ₹50,000 to ₹1,00,000 per month, depending on the prior profile and performance in our assessments.

If your strengths are broader than hands-on research, for example research support, operations, or communications, the Special Projects Fellow role is likely a better fit.

How to Apply

Please read the complete job description before applying. We review applications as they arrive, and we encourage you to apply early using the Apply Now button on this page.

Location

Hybrid. You will start remote for the first month. Eventually, we may require you to be located in one of our Secure AI Hubs, or elsewhere across India.

Type

Full-time fixed-term contractor role, initially for 6 to 12 months, with considerable autonomy. The first two months are a probation period. Based on satisfactory or outstanding performance, there are reasonable prospects to renew the contract or to graduate into a full-time staff researcher role at the lab. You will work closely with SAFL's research leads and leadership, and with academic and international partners in other time zones.

Compensation (INR)

₹1,00,000 to ₹4,00,000 per month. Compensation is all-inclusive and will be based on your prior relevant work and your demonstrated ability to meet the needs of this role. For people with significant relevant experience, we can also hire for a senior version of this role.

Application Process Overview

Our process may change based on our needs, but here is what you can usually expect.

• Screening application. A short application to check basic fit.
• Skills test. An unpaid 2-hour work test to assess relevant skills.
• Video interview (optional). A short automated video interview of about 5–7 minutes to help us understand you.
• Interviews. 1–2 short interviews. These go both ways. We learn more about you and you learn more about the role and about us.
• Paid work trial. A paid trial of 2–5 days where you work on real tasks with the team, in person or remote.
• Final call. A closing conversation before we decide.
• Reference checks. We will reach out to a couple of your references to wrap things up.

SAFL is committed to building a diverse, curious, and high-integrity team. We welcome applications from people of different backgrounds.

Questions? Write to us at hiring[AT]secureaifutureslab[DOT]com.