Member of Technical Staff - Research Software Engineer
Reflection5 months ago
Responsibilities
- Design and optimize large-scale reinforcement learning training loops and data pipelines.
- Implement state-of-the-art techniques while ensuring numerical stability and computational efficiency.
- Build internal tooling for launching, monitoring, and reproducing complex experiments.
- Diagnose GPU memory issues, communication overhead, dataloader stalls, and other training-stack bottlenecks.
- Translate research prototypes into reusable, production-grade infrastructure.
- Architect and optimize distributed GPU training systems, inference systems, and core model-training infrastructure.
Requirements
- Strong software engineering skills and fluency with machine learning concepts.
- Ability to implement research papers and translate research ideas into reliable systems; a PhD is not required.
- Deep experience in at least one of distributed training and inference or data infrastructure.
- Experience working across machine learning algorithms, distributed systems, and high-performance computing.
- Strong focus on performance, numerical stability, and reproducibility.
Benefits
- Top-tier compensation and equity, with no specific base salary amount stated.
- Comprehensive medical, dental, vision, life, and disability insurance.
- Fully paid parental leave for all new parents, including adoptive and surrogate journeys.
- Financial support for family planning.
- Paid time off and relocation support.
- Daily lunch and dinner, regular off-sites, and team celebrations.
Tech Stack
Categories
About Reflection
Reflection is a research lab making intelligence open and accessible for everyone to use, customize, and build on. Our team previously built frontier LLMs at labs like DeepMind, OpenAI, and Anthropic. We believe AI should be built in the open, with transparent research and collaborative development. That means giving enterprises, governments, and sovereign entities true ownership and control of AI that performs at the highest level. Our mission: make intelligence open and accessible to all.