Research Scientist

about the company

Our client is building the next generation of specialized, efficient large language models. Founded by alumni from Google Research, Apple, Stanford, and Cambridge, the team is focused on developing high-performance AI systems that outperform traditional LLM benchmarks in targeted domains. Their open-source model family has surpassed 5 million downloads and is used by companies including NVIDIA, Meta, and Airbnb. The company has raised $25M in seed funding, backed by leading investors.

what we're looking for

We're seeking a Research Scientist – Small Language Models to drive architectural innovation and advance the research roadmap behind next-generation compact models. This role is ideal for a deeply technical researcher with strong experience in large-scale deep learning who can bridge cutting-edge experimentation with production impact.

responsibilities

• Research and experiment with novel language model architectures
• Optimize multimodal models to improve instruction-following, response quality, and performance metrics
• Design and implement large-scale data processing pipelines, including filtering, balancing, and captioning systems
• Apply reinforcement learning techniques (e.g., DPO, GRPO) to align model outputs with human preferences
• Build robust, real-world evaluation frameworks
• Collaborate with engineering to ship research advancements into production systems
• Establish strong standards for experimentation, reproducibility, and documentation

qualifications

• Master's or PhD in Computer Science, AI, Machine Learning, or related field
• Demonstrated ability to conduct independent research in academic or industry settings
• Significant experience training large-scale deep learning models
• Expertise in at least one of: Large Language Models, Vision-Language Models, Diffusion Models, or comparable generative architectures
• Strong proficiency in leading deep learning frameworks (PyTorch, JAX, TensorFlow)