Member of Technical Staff, Distributed Training Performance

Hark

San Jose (CA)

On-site

USD 180,000 - 450,000

Full time

9 days ago

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Job summary

Hark is building next-generation AI systems and hardware to enable multimodal, real-time interactions across speech, text, vision, and memory. The Omni team extends Megatron-based pretraining across thousands of GPUs to deliver scalable, high-throughput training.

As a key member, you will optimize MFU, implement fused kernels, and advance model and data parallelism for dense and mixture-of-experts architectures, while ensuring numerical stability and reliable large-scale runs.

Qualifications

  • Experience with large-scale distributed training systems (Megatron-LM/NeMo or similar).
  • Strong understanding of data, tensor, pipeline, sequence, context, and expert parallelism.
  • Experience using GPU profilers to diagnose performance bottlenecks.
  • Proficient in Python and C++/CUDA with tooling for researchers.
  • Experience with mixed-precision training and numerical stability.

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 and routing.
  • Overlap communication with compute to raise throughput per GPU.
  • Implement model and optimizer sharding and optimize various parallelism forms.
  • Profile GPU kernels and identify performance bottlenecks; ensure numerical stability.
  • Develop reusable training components for multimodal encoders (speech, vision, text).
  • Investigate training failures and coordinate with researchers and infra teams.
  • Improve checkpointing, restart time, and observability; participate in on-call support.

Skills

Megatron-LM familiarity
Python & C++/CUDA
CUDA kernels
GPU profiling
Distributed training
Mixed-precision training
Model/optimizer sharding
Multimodal training

Tools

Triton
CUTLASS
NCCL
Slurm
Kubernetes

Job description

About Hark

Hark is an artificial intelligence company building advanced, personalized intelligence. One that is proactive, multimodal, and capable of interacting with the world through speech, text, vision, and persistent memory.

We're pairing that intelligence with next-generation hardware to create a universal interface between humans and machines. While today's AI largely operates through chat boxes and decade-old devices, Hark is focused on what comes next: agentic systems that interact naturally with people and the real world.

To get there, we're developing multimodal models and next-generation AI hardware together - designed from the ground up as a single, unified interface for a new era of intelligent systems.

About the Role

The Omni team at Hark is building the next generation of AI experiences beyond text, enabling models to understand and generate content across multiple modalities, including text, audio, and beyond. Our goal is to create seamless, real-time multimodal intelligence that powers intuitive and immersive user experiences.

As part of the Omni team, you'll improve the performance of Hark’s Megatron-based pretraining systems across thousands of GPUs. The role focuses on model FLOPs utilization (MFU), training throughput, and the performance of kernels, communication, and parallelism strategies. You'll work with researchers to support new architectures and with infrastructure engineers to resolve bottlenecks, maintain numerical stability, and keep large runs reliable.

Responsibilities
  • Own MFU and step time for pretraining runs across dense and mixture-of-experts models, setting targets per model size and tracking them across runs.
  • 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.
  • Develop reusable training components that support new architectures and multimodal encoders for speech, vision, and text.
  • Investigate performance regressions and training failures involving communication, numerical changes, hardware faults, or straggling nodes.
  • Work with architecture researchers and compute infrastructure engineers to align model designs, training software, and cluster capabilities.
  • Improve checkpointing, restart time, failure detection, and observability to reduce lost training work. Participate in on-call support during large runs.
Requirements
  • Experience developing distributed training systems and working with the internals of Megatron-LM, NeMo, or a comparable framework, including its parallelism strategies.
  • Strong understanding of data, tensor, pipeline, sequence, context, and expert parallelism, including communication patterns, memory tradeoffs, and failure modes.
  • Experience using GPU profilers to analyze kernel execution, memory bandwidth, communication overhead, and data-loading stalls.
  • Working knowledge of mixed-precision training, including BF16 and FP8, and methods for diagnosing numerical instability.
  • Strong Python and C++/CUDA engineering skills, with experience building reliable tools for researchers.
  • Experience developing CUDA or Triton kernels, or using CUTLASS, for attention, normalization, or mixture-of-experts models.
  • A record of measurably improving MFU or throughput on large-scale training jobs.
Bonus Qualifications
  • Experience with mixture-of-experts training at scale, including expert parallelism, load balancing, and all-to-all communication optimization.
  • Experience with GPU communication libraries such as NCCL or network topology-aware scheduling.
  • Experience with fault-tolerant training, asynchronous checkpointing, in-memory recovery, or elastic training.
  • Experience balancing workloads across different encoders in multimodal training.
  • Familiarity with cluster schedulers (Slurm, Kubernetes) and GPU health monitoring at fleet scale.
Compensation

The US base salary range for this full-time position is between $180,000 - $450,000 annually.

The pay offered for this position may vary based on several individual factors, including job-related knowledge, skills, and experience. The total compensation package may also include additional components/benefits depending on the specific role. This information will be shared if an employment offer is extended.

Get your free, confidential resume review.

or drag and drop your file here.

Similar jobs

Similar jobs worth comparing

Member of Technical Staff, Distributed Training Performance San Jose
Member of Technical Staff, Distributed Training Performance San Jose

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Pretraining Architecture and Scaling
Member of Technical Staff, Pretraining Architecture and Scaling

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Pretraining Architecture and Scaling San Jose
Member of Technical Staff, Pretraining Architecture and Scaling San Jose

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Mid-training San Jose
Member of Technical Staff, Mid-training San Jose

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Pretraining
Member of Technical Staff, Pretraining

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Mid-training
Member of Technical Staff, Mid-training

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Pretraining Evaluations
Member of Technical Staff, Pretraining Evaluations

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Multimodal
Member of Technical Staff, Multimodal

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Multimodal Pretraining Data San Jose
Member of Technical Staff, Multimodal Pretraining Data San Jose

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000
Member of Technical Staff, Multimodal Vision
Member of Technical Staff, Multimodal Vision

Hark • San Jose (CA)

On-site
USD 180,000 - 450,000