Improve the performance of Hark's Megatron-based pretraining systems across thousands of GPUs, focusing on model FLOPs utilization, training throughput, and kernel performance.
Compensation
The US base salary range for this full-time position is between $180,000 - $450,000 annually.
Responsibilities
Own MFU and step time for pretraining runs across dense and mixture-of-experts models
Write or integrate fused and custom kernels (CUDA, Triton, CUTLASS) for attention, normalization, and mixture-of-experts routing
Overlap communication with compute, reduce pipeline bubbles, and tune activation recomputation and memory use to raise throughput per GPU
Implement model and optimizer sharding, and optimize data, tensor, pipeline, sequence, context, and expert parallelism for dense and mixture-of-experts architectures
Profile GPU kernels, collective communication, data loading, and mixed-precision training to identify performance and numerical stability bottlenecks
Requirements
Nice to have: mixture-of-experts training at scale
Nice to have: load balancing
Nice to have: all-to-all communication optimization
Apply AI engineering expertise to deploy, optimize, and scale models for strategic customers, leading technical journeys from POCs to production endpoints.
Responsibilities
Scope workloads, define metrics (TTFT, throughput, cost/token), build benchmarks, and lead technical POCs.
Serve as the technical lead for customer engineering teams through architecture reviews and rollout planning.
Tune engines (KV cache, speculative decoding, parallelism, quantization) to hit aggressive performance targets.
Profile and diagnose bottlenecks at the engine level to deliver production fixes.
Own end-to-end customer onboarding, deployment, autoscaling, capacity planning, and SLO monitoring.
Build and own full-stack features for Sentry’s Size Analysis and Snapshots products, from backend pipelines to frontend UI, ensuring efficient handling of large mobile build artifacts and visual diffing.
Compensation
We are obliged by law to quote the minimum wage of EUR 60,844,-- gross per year for this position, in accordance with the respective collective agreement (IT-KV). The salary that Sentry offers you is adjusted to the current market situation and will be above the IT-KV.
Responsibilities
Build and tune backend services and task-based pipelines that ingest, analyze and compare builds
Extend product features end-to-end and propose further development
Own GitHub integration end-to-end, including PR status checks and webhooks
Ship frontend features such as treemaps, build comparisons and diff viewers
Deploy across regions, monitor metrics, logging, and handle on-call duties
Help make the latest AI models run well on our silicon by focusing on LLM inference, numerical optimization, and building agentic workflows for model implementation.
Locations (4)
Sunnyvale, California, United States
Munich, Bavaria, Germany
North America, United States
Europe
Responsibilities
Build agentic workflows to give agents tools to inspect models, change implementations and numerical settings, run emulation and hardware experiments, and diagnose failures.
Bring up new models quickly and work with engineers across the stack to take architectures from reference code to performant execution.
As a Silicon Architecture Graduate Intern, you will contribute to the definition and architecture of cutting‑edge silicon solutions, including hardware, firmware, and platform specifications.
Compensation
Annual Salary Range for jobs which could be performed in the US: $105,698.00-105,702.00 USD
Responsibilities
Assist in defining silicon architecture specifications based on product requirements and feature integration.
Support modeling and simulation tasks to evaluate performance, power, and trade‑offs for architectural decisions.
Collaborate with architects, design engineers, and validation teams to ensure feasibility and alignment with product goals.
Contribute to the development of high‑level models and tools for architectural analysis.
Explore GPU concepts such as SIMT, parallelism, and cache hierarchies to guide architectural improvements.
Lead and contribute to core infrastructure and shared functionality for Elasticsearch, ensuring performance, maintainability, and security across cloud, serverless, and on-premise deployments.
Locations (4)
Seattle, Washington, United States
Los Angeles, California, United States
San Francisco Bay Area, California, United States
New York City, New York, United States
Compensation
The typical starting salary range for this role is: $133,100—$210,600 USD. The typical starting salary range for this role in the select locations listed above is: $159,900—$252,900 USD
Responsibilities
Contribute to large, impactful projects that evolve the core infrastructure that underpins Elasticsearch
employee benefit plans/programs applicable to the candidate’s position (including incentive compensation, equity grants, paid time off, and group health insurance coverage)
Member of Technical Staff, Distributed Training PerformanceNewHark · San Jose, US
Staff Applied AI EngineerNewCrusoe · Tel Aviv, Israel
Software Engineer, Size Analysis & SnapshotsSentry · Vienna, Austria
Senior Software Engineer — ML Inference & NumericsTensordyne · Remote, US+3 more
Sr. Software Engineer, Training/Inference (C++)SpaceXAI · Palo Alto, US
Senior Java Engineer - Core/Infra - ElasticsearchElastic · Seattle, US+3 more