ErgodicAbout this role
Summary: Fixed-term PhD fellowship for students or graduates to develop autonomous agents and world models using LLMs, active inference, or optimal policies. The role offers real-world problem-solving experience within an early-stage AI startup.
Company Description
Ergodic AI is an AI startup based in London that is building an action-oriented AI platform to help enterprises move from analysis paralysis to rapid data-driven execution.
Kepler - our AI platform - identifies context, surfaces optimal actions and justifies the reasons for their execution. We're an early stage startup and are already working with multiple F500 enterprises.
We’re backed by leading VC funds to help further our mission and are now on the hunt for the brightest and best to join us.
Our Team
The founding team is highly experienced across a range of AI and commercial work, with careers in world class organisations and experience building and exiting $multi-million startups previously.
Process
As you progress through the process, you’ll meet each of the 3 founders who will probe you for your motivations, experiences and fit for the role.
Steps:
- Application / CV screening
- Interview with CTO
- Interview with CSO
- Interview with CEO
- Decision
We aim to move quickly through the process and not waste your time.
The Role:
This is fixed-term role for PhD students or PhD graduates looking to transition to an industry role. It lasts for either 3 or 6 months and gives you the ability to experience real-world problems, while still researching cutting edge AI.
Our research plan is to develop autonomous agents that can create world models - internal representations of the environment - and navigate through them. As an AI Researcher, you will be expected to develop models in one of the spaces:
- Large language models - experimenting with different architectures and fine-tuning existing solutions.
- Active Inference - building general purpose deep-learning based agents that interact with the environment and learn from these interactions.
- Optimal Policies - based on a series of possible actions, what is the best one?
We have offices in London and Munich, and accept remote candidates as well.
Qualifications
- MSc in a quantitative field, either working on a PhD or have finished the PhD
- Familiarity with standard ML libraries (pytorch, sklearn, etc..)
- Hacker mentality: there's always to a way to solve a problem.
- Ability to break down a problem in simple steps.
- Demonstrable track record of expertise/resourcefulness: either articles, open-source contributions, interesting projects