About the Role
Peritium Search have partnered with a stealth AI company building next-generation intelligent systems at the frontier of applied machine learning. The team is focused on transforming what they believe is a nuanced intersection of AI research.
As a Machine Learning Researcher specialising in Large Language Models (LLMs), you will help shape the research direction and technical foundation of our core AI platform. You will work across the full model lifecycle—from experimentation and evaluation to optimisation and deployment, tackling challenging real-world AI fundamentals that demand robust reasoning, reliability, and scalability.
This is an opportunity to work with a small, highly technical team all of whom have worked at elite Frontier Labs across the entire training spectrum with huge success. I should note to any perspective candidate, that your research WILL directly influence product direction and have immediate impact.
What will you do/build
- Design, build, and optimise LLM training, fine-tuning, and evaluation pipelines.
- Research and implement post-training techniques including supervised fine-tuning (SFT), Direct Preference Optimisation (DPO), reinforcement learning from human feedback (RLHF), and emerging alignment methods.
- Develop novel approaches to improve reasoning, factuality, robustness, and reliability in production AI systems.
- Build retrieval-augmented generation (RAG) pipelines and embedding-based systems that integrate external knowledge sources.
- Design rigorous evaluation frameworks, benchmarks, and automated testing for model quality and safety.
- Collaborate closely with engineers and product leaders to translate research into scalable, production-ready capabilities.
- Stay at the forefront of advances in foundation models, reasoning systems, reinforcement learning, and agentic AI.
Key Role Requirements
- Strong background in machine learning, deep learning, reinforcement learning, and optimization.
- Hands-on experience with LLM post-training techniques including SFT, DPO, RLHF, reward modelling, or related methods.
- Strong Python skills and experience with frameworks such as PyTorch, JAX, or TensorFlow.
- Experience working with large-scale datasets, distributed training, and GPU optimisation.
- Familiarity with retrieval systems, embeddings, vector databases, and RAG architectures.
- Ability to navigate and then thrive in an early-stage 0-1 environment where priorities evolve quickly and ownership is expected.
- Strong communication skills and the ability to move rapidly from research ideas to working systems.
Nice to Have
- Experience with multimodal models, AI agents, or reasoning-focused LLMs.
- Contributions to open-source ML projects or publications in machine learning or AI.
- Experience with distributed systems, inference optimisation, or model serving.
- Knowledge of quantitative domains, complex decision-making systems, or high-performance computing.
Why Join
- Work on frontier AI research with the opportunity to shape the direction of an ambitious stealth company.
- Build products from first principles alongside a highly experienced technical founding team.
- High levels of ownership, autonomy, and technical influence from day one.
- Competitive salary, meaningful equity (this role really does have meaningful equity), and the opportunity to help define the future of the company.