What you'll do
- Survey and synthesize state-of-the-art research across LLMs, vision AI, and agentic systems
- Reproduce key papers and adapt techniques for production use cases
- Run experiments to evaluate model architectures, training strategies, and inference optimisations
- Contribute to internal knowledge base and publish technical blog posts
- Collaborate with engineering teams to transition research into deployable systems
What we're looking for
- Strong mathematical foundations — linear algebra, probability, statistics, and calculus
- Hands-on experience with PyTorch or JAX
- Ability to read, implement, and critique ML research papers
- Familiarity with fine-tuning workflows (LoRA, QLoRA, RLHF, DPO)
- Excellent written English for documenting findings
Nice to have
- Publications or pre-prints (NeurIPS, ICML, ACL, ICLR, or arxiv)
- Experience with distributed training frameworks
- MSc or PhD in CS, ML, or a related quantitative field
This is a research-first role for someone who genuinely loves reading papers and turning insights into working systems. You’ll influence the technical direction of Superteams and directly impact what our clients ship.