
Principal Engineer: Annapurna Labs Training and Inference, Annapurna Labs, AWS Neuron SDK
Amazon · Posted Oct 14, 2026
- Seattle, Washington, United States +1 more
- On-site
- Full-time · Principal
- $200K–$311K per year
- 10+ years
Principal Engineers are both technical leaders and hands-on builders, working on the AWS Neuron SDK team within Annapurna Labs to drive AI model performance across the lifecycle on AWS Trainium — pre-training, post-training and alignment, and inference serving.
Locations (2)
- Seattle, Washington, United States
- Cupertino, California, United States
Compensation
USA, CA, Cupertino - 230,100.00 - 311,200.00 USD annually USA, WA, Seattle - 200,100.00 - 270,600.00 USD annually
Responsibilities
- Kernel optimization for ML training — forward and backward path — and drive performance out of it, with training collective algorithms (collectives, all-reduce, cross-accelerator coordination)
- Prove top-tier performance even at small scale, and clear the path so customers can land AI models on the chip out of the box
- Deliver first-class RLHF/RLAIF and preference methods (PPO, GRPO, DPO/IPO/KTO, reward modeling) with training-plus-inference co-location and convergence safeguards (KL control, reward-scaling, policy-collapse detection), and native performance for the post-training ecosystem (TRL, veRL, NeMo-Aligner, TorchTitan/TorchAO, JAX)