Research Engineer, Post-Training
San FranciscoFull-timePosted 11mo ago
About the role
We are an applied AI lab building end-to-end software agents. Our team develops Devin, the first AI software engineer, and seeks a Research Engineer focused on post-training methods to bridge raw model capability with real-world, safe, and effective systems.
Responsibilities
- Post-Training Recipe Development: Iterate on the full stack of datasets, training stages, and hyperparameters that determine model behavior. Measure how choices compound across evals and production performance, not just isolated benchmarks.
- Evaluation Design and Integrity: Build evals that actually capture what matters. Define, optimize, realize the gaps, and rebuild to improve numbers meaningfully.
- Deep Understanding: Investigate results that don’t make sense and carry insights forward to future problems.
- Alignment and Agent Behavior: Apply and advance techniques like RLHF, RLAIF, and constitutional approaches to shape agent reasoning, actions, and collaboration with humans in long-horizon tasks.
- Scaling and Exploration: Measure performance scaling with data and compute; develop new methodologies when existing ones reach limits.
Qualifications
- A track record of advancing ML systems through post-training, alignment, or related methods: RLHF, RLAIF, preference modeling, reward learning, or equivalent
- Strong fundamentals in probability, statistics, and ML theory; ability to interpret experimental data and distinguish real effects from noise and bugs
- Evidence of original contributions: publications at top venues, open-source impact, or equivalent industry results
- Experience with large-scale distributed training and debugging
- Systems-level thinking: understanding how training pipelines, data, and evaluation interact
- Comfort with ambiguity and fast-moving research environments where priorities shift quickly
- PhD is not required; considered as one signal among many
Skills
Required:Post-TrainingEvaluation Design
Nice to have:RLHF/RLAIF/Constitutional Approaches
Benefits
- Compute is not a constraint: large allocations with training jobs across thousands of GPUs from day one
- Environment rewards speed, autonomy, and technical depth with minimal process overhead
- Prototypes reach real deployment quickly
- Everything needed to operate at frontier scale from day one
Last updated Oct 21, 2026. Interested? Apply on the company site.
About Cognition
cognition is an applied AI lab
An applied AI lab building the future of software engineering. We’re the makers of Devin, a collaborative AI teammate that helps ambitious engineering teams achieve more.
AIai-plg-top-20collaborative AIliquid-ai-recruiting-targetmachine learningsoftware engineering
More open roles at Cognition
- Sales Development Manager5d agoSan Francisco, Austin, New York City · Sales
- Partner Director (LATAM)5d agoSao Paulo · Sales
- Analyst Relations5d agoSan Francisco · Marketing
- Head of Safety1w agoSan Francisco · Research & Development
- Japan Partner Marketing1w agoTokyo · Marketing