AfterQuery
Research Engineer
About this role
AfterQuery seeks a Research Engineer to design and operate training experiments measuring how datasets impact foundation model behavior. You'll build post-training infrastructure for SFT, RL, and evaluation while orchestrating reproducible experiments that isolate the effects of different data strategies on model performance and alignment.
What you'll do
- Design and build post-training infrastructure for SFT, RL, and evaluation workflows
- Create reliable data preparation, versioning, sampling, and checkpoint management pipelines
- Develop experiment orchestration and tracking systems ensuring reproducibility and easy debugging
- Build reusable abstractions enabling rapid experiment launches across datasets and models
- Integrate systems with partner-lab training stacks, APIs, and evaluation infrastructure
- Formulate hypotheses, run models, analyze results, and identify confounders to drive next steps
What they're looking for
- Python and software engineering fundamentals
- PyTorch and JAX frameworks
- Ray and Slurm distributed computing
- LLM fine-tuning, post-training, and evaluation
- ML training and data infrastructure development
- Experimental design and statistical analysis
- Debugging distributed systems and training pipelines
- Reproducible ML systems architecture
Benefits
- Medical, vision, and dental insurance
- 401(k) with employer match
- Daily meal stipend via UberEats
- Monthly wellness stipend covering Equinox membership
- Commute coverage
- Meaningful equity package
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
AfterQuery
AfterQuery is an applied AI research lab that builds data infrastructure and evaluation frameworks powering foundation model development for frontier AI labs. The company is hiring fullstack software engineers, infrastructure/security specialists, and interns to design scalable data pipelines, develop datasets and reward signals, and create systems that directly influence how advanced AI models are trained.
View all jobs at AfterQueryLikely interview questions
- Describe a time you built a training pipeline or infrastructure that handled multiple experiments—how did you ensure reproducibility and easy comparison?
- Walk us through your experience with LLM fine-tuning and post-training. What were the biggest challenges and how did you address them?