SID trains AI that can retrieve and reason over any data source. Intelligence and skills are inconsequential without context. Today, AI is blind to information that is not on the internet. If we want AI to solve real problems, we need to change that.
Requirements
- Familiar with RL pipelines for language models
- Comfortable with torchrun/accelerate/multi-node training.
- Clever about getting the data needed – or synthetically generating it.
- Finds easy solutions to hard problems, but doesn't mind getting their hands dirty, i.e., jumping a layer down into PyTorch or CUDA.
- Familiar with 'You and Your Research.' Understands what it takes to do significant work.
Responsibilities
- Post-train reasoning into LLMs with GRPO and SFT.
- Design and iterate RL training environments for retrieval – unstructured, structured, web.
- Run small and large model experiments – yolo runs encouraged.
- Work on next-generation vision-first embedding models.
Other
- US citizen/visa only
- Not afraid of formulas – a technical major is an indicator of this (but isn't the only one).
- Thinks they can learn anything in 2 weeks, but isn't arrogant about it.
- Prefers .py to .tex
- Must articulate ideas well! A big part of making successful models is telling people about them. This includes writing docs and technical reports at the minimum – and jumping on podcasts at the extreme.
