This role pays more than
88%
of similar roles. Most pay
$167,449–$248,375
— the shaded band above.
At the midpoint, this role pays about
$265k
versus about
$208k
for comparable roles.
Based on 240 similar postings.
Employer
About Adobe
Adobe Inc. is a global software company known for creative and multimedia software products including Photoshop, Illustrator, Acrobat, and its cloud-based Creative Cloud and Document Cloud suites. Industry: Creative & Digital Experience Software
Adobe currently has
218 open roles
on FindRole.
Listed pay typically runs
$187,100–$270,950
across 216 roles with salary data.
Senior Applied Scientist / Engineer, Training & Inference joins the Applied Science & Machine Learning team to bridge the gap between research and production for next-generation video and multimodal generation models. This role serves as a technical owner for the training-to-deployment pipeline, ensuring large generative models are delivered reliably, performantly, and cost-efficiently. Key responsibilities include implementing distributed training strategies like PyTorch FSDP, Tensor Parallelism, and Pipeline Parallelism across multi-node GPU environments, alongside designing inference serving systems focused on latency and throughput. The candidate will utilize Python and PyTorch to harden, validate, and operationalize models at scale. Preferred experience includes working with diffusion models or flow matching, as well as familiarity with frameworks like TensorRT or vLLM. This role solves the critical challenge of transitioning large-scale model checkpoints into stable production environments while optimizing for GPU efficiency and cost.
Own the end-to-end training-to-deployment pipeline for video and multimodal generative models.
Implement distributed training strategies using PyTorch FSDP, Tensor Parallelism, and Pipeline Parallelism on multi-node GPU environments.
Design and optimize inference systems to improve latency, throughput, and cost efficiency across production targets.
Bridge the gap between research checkpoints and reliable production by hardening and validating models at scale.
Identify and resolve inefficiencies in memory, communication, and execution orchestration to meet GPU performance goals.
Develop and maintain production-critical machine learning systems using Python and PyTorch.
What we're looking for
Master's or PhD in Computer Science, Electrical Engineering, AI/ML, or a related field, or equivalent practical experience.
Hands-on experience with large-scale distributed training using PyTorch (FSDP, Tensor Parallelism, Pipeline Parallelism) across multi-node GPU environments.
Proven experience optimizing and deploying large generative models for production, including serving infrastructure and cost-aware deployment.
Proficiency in Python and PyTorch within large shared codebases and production-critical ML systems.
Demonstrated ability to lead end-to-end technical areas and navigate the engineering challenges of reliability and scale.
Experience training and deploying video, image, or multimodal generative models (preferred).
Familiarity with inference serving frameworks such as TensorRT, vLLM, or equivalent (preferred).
Experience with performance profiling and optimization for both training and inference workloads (preferred).