Research Scientist, Interpretability

Research · Full-time

Remote (US / EU overlap)

Research Scientist, Alignment

Research · Full-time

Remote (US / EU overlap)

Research Engineer, Evaluations

Safety & Evaluations · Full-time

Remote

Senior Machine Learning Engineer, Serving

Engineering · Full-time

Remote

Machine Learning Engineer, Training Infrastructure

Engineering · Full-time

Remote

Applied AI Engineer, Product

Product Engineering · Full-time

Remote

Policy Researcher, AI Governance

Policy & Governance · Full-time

Remote

Program Lead, Community & Funding

Operations · Full-time

Remote

Operations & Finance Manager

Operations · Full-time

Remote

Research Scientist, Interpretability

ResearchRemote (US / EU overlap)Full-time

Join the team working to understand what our models are actually doing — the circuits, features, and mechanisms behind their behavior. Your work will feed directly into the safety evaluations we publish before every release.

What you'll do

  • Design and run experiments on internal representations of Krus, Mantus, and Corus
  • Build tools that let other researchers and reviewers inspect model behavior
  • Publish findings openly and write for both research and general audiences
  • Work with the evaluations team to turn interpretability findings into testable commitments

What we're looking for

  • A track record of rigorous research in machine learning, preferably on interpretability or related areas
  • Comfort with large-scale experiment tooling and analysis
  • A belief that interpretability work matters most when it changes what gets deployed

This role would report to Lucas Ferrand, Head of Research.

Research Scientist, Alignment

ResearchRemote (US / EU overlap)Full-time

Help us make Ichnus genuinely helpful, honest, and honest about its limits. This is applied alignment work — the methods we develop go into production models used by millions of people.

What you'll do

  • Develop and test training and post-training methods that improve helpfulness and honesty
  • Study when models overstate their confidence, and build methods that reduce it
  • Contribute to the "deference under uncertainty" research line and its product behavior
  • Collaborate with the evaluations team to measure alignment improvements

What we're looking for

  • Strong background in deep learning and practical alignment or RLHF-style work
  • Experience shipping model behavior changes into a production product
  • Comfort moving between research questions and product constraints

This role would report to Priya Raman, Head of Alignment.

Research Engineer, Evaluations

Safety & EvaluationsRemoteFull-time

Build the infrastructure that tests every Ichnus model before it reaches people — capability benchmarks, safety evals, and adversarial red-teaming at scale.

What you'll do

  • Own the evaluation harness that gates model releases
  • Design new evals where existing ones are not good enough, including public-facing safety evals
  • Run and analyze red-teaming campaigns with internal and external reviewers
  • Publish evaluation results and methodology openly

What we're looking for

  • Strong engineering skills in Python and distributed evaluation tooling
  • Experience with LLM evaluation or a related testing-heavy discipline
  • Comfort with adversarial thinking and a low tolerance for "good enough"

This role would report to Tomás Herrera, Head of Evaluations.

Senior Machine Learning Engineer, Serving

EngineeringRemoteFull-time

Keep Ichnus fast, reliable, and affordable for everyone. Your systems are what make "free by default" economically possible at scale.

What you'll do

  • Build and operate the serving stack behind Krus, Mantus, and Corus
  • Work on routing, quantization, and tiered compute allocation
  • Keep latency and reliability targets while controlling cost
  • Pair with the efficiency research team to ship new methods to production

What we're looking for

  • Deep experience with production ML serving systems (inference engines, GPU clusters)
  • Strong systems engineering fundamentals
  • Comfort with on-call for critical infrastructure

This role would report to Hana Sato, Engineering Lead.

Machine Learning Engineer, Training Infrastructure

EngineeringRemoteFull-time

Own the training runs behind every model we release — from data pipelines to distributed training to the tooling that makes runs reproducible.

What you'll do

  • Build and maintain distributed training infrastructure
  • Own data pipelines for pre-training and post-training data
  • Instrument runs so researchers can move fast without breaking things
  • Document everything so future runs are easier than the last

What we're looking for

  • Experience running large-scale ML training workloads
  • Strong systems programming and Python skills
  • A love of boring, reliable infrastructure

This role would report to Hana Sato, Engineering Lead.

Applied AI Engineer, Product

Product EngineeringRemoteFull-time

Turn research and model behavior into the Ichnus people actually use — the assistant, its apps, and the experiences that make it feel like something you can trust.

What you'll do

  • Build and maintain the Ichnus apps and their interaction design
  • Prototype features that surface model capability and limits honestly
  • Work with the community team to incorporate user feedback into the product
  • Keep quality high: we would rather ship less than ship something we cannot stand behind

What we're looking for

  • Strong product engineering experience (web and mobile)
  • Design sensibility and attention to detail
  • Experience building with LLMs a plus, a willingness to learn a must

This role would report to Hana Sato, Engineering Lead.

Policy Researcher, AI Governance

Policy & GovernanceRemoteFull-time

Study how AI should be governed — and help Curos live up to the standards we argue for. Your work spans research papers and the internal charters we are held to.

What you'll do

  • Research questions of accountability, transparency, and public benefit in AI
  • Draft and maintain Curos's governance documents and release commitments
  • Liaise with the board and oversight committees
  • Publish policy research openly and in plain language

What we're looking for

  • A research or practitioner background in AI policy, governance, or law
  • Strong writing, including for non-expert audiences
  • Comfort with the tension between ideals and operational reality

This role would report to Grace Adeyemi, Director of Policy & Governance.

Program Lead, Community & Funding

OperationsRemoteFull-time

Run the two relationships that keep Curos independent: the community that uses Ichnus, and the donors and foundations that fund it. You will own our community council and our public funding reports.

What you'll do

  • Run the community council — elections, meetings, and follow-through
  • Manage grant applications and relationships with foundations
  • Maintain and publish the transparency reports with the ops team
  • Run community events, from office hours to public red-teaming sessions

What we're looking for

  • Excellent project management and stakeholder communication
  • Experience with community programs or grant work
  • A comfort with the public nature of our commitments

This role would report to Sam Whitfield, Head of Operations.

Operations & Finance Manager

OperationsRemoteFull-time

Keep the finances boring and transparent. You own budgeting, accounting, and the numbers that make our transparency reports possible.

What you'll do

  • Own budgeting, forecasting, and spend tracking
  • Manage the annual audit and financial reporting
  • Build the numbers behind the transparency page
  • Keep costs lean without cutting into mission-critical work

What we're looking for

  • Strong finance and accounting skills (non-profit experience a plus)
  • Meticulous, honest, and comfortable with public scrutiny of the numbers
  • Enthusiasm for boring, correct finance

This role would report to Sam Whitfield, Head of Operations.

How we hire

No take-home that takes a weekend. No whiteboard trivia. The process is a conversation, a short paid exercise, and a chance to talk with the team you'd join. We care most about whether you can do the work and believe in the mission.

We pay a fair, transparent salary band for every role — no negotiation roulette, no equity, because there is no equity to give. Health care, retirement, and generous leave are standard.

We are not currently hiring. Curos is a concept project, and these listings are fiction. If you love the idea of non-profit AI, we hope you'll build it — or join a real organization working on it.

Ask us something anyway →