Member of Technical Staff, Multimodal Vision

HarkSan Jose AI Foundation Models

Apply on company site

Opens the official Hark careers page

Salary

$180K - $450K USD

Job type

Full-time

Posted

23 Mar 2026

Status

Open · checked 7d ago

About this role

Hark is currently seeking a Member of Technical Staff, Multimodal Vision to join their team in San Jose. This role suits an artificial intelligence professional with a strong background in multimodal foundation models, vision-language systems, and advanced training infrastructure.

Role Overview

The Omni team at Hark is shaping the future of artificial intelligence by developing systems that go beyond traditional text. Professionals in this position will work on real-time multimodal intelligence, handling everything from foundational model research to product integration. The objective is to create immersive experiences that combine text, vision, and audio into a unified interface.

Key Responsibilities

Advance multimodal foundation models through mid-training and post-training research using large-scale vision-text and video-audio data.

Take charge of the reinforcement learning stack, covering reward modeling, policy optimization, preference optimization, and mitigation of capability regression.

Design comprehensive multimodal evaluation frameworks and internal benchmarks to measure perception, reasoning, and instruction following.

Optimize training infrastructure, focusing on distributed training, throughput, memory efficiency, and scalable rollout systems.

Requirements

Demonstrated history of improving multimodal or vision-language models through novel contributions in data, training systems, or model tuning.

Extensive hands-on experience with large-scale post-training methods like supervised fine-tuning, reward modeling, and preference optimization.

Solid practical knowledge in video modeling, image understanding, and multimodal machine learning.

Proven expertise in data-driven experimentation, evaluation design, and fast-paced iterative model development.

Nice to Have

Practical background in scaling reinforcement learning training infrastructure.

Experience working with speech-language models, video-audio systems, or streaming real-time multimodal technologies.

Familiarity with large-scale distributed machine learning frameworks like DeepSpeed, FSDP, or Megatron-LM.

Documented open-source contributions or research publications at top-tier artificial intelligence venues.

Skills and Tools

Multimodal Learning

Vision-Language Models

Reinforcement Learning (PPO, DPO, GRPO)

Supervised Fine-Tuning

Distributed Training Infrastructure

Salary and Benefits

The annual base salary range for this position spans from $180,000 to $450,000 in the United States. Final compensation offers depend on individual qualifications, knowledge, and relevant experience.

About the Company

Hark is an artificial intelligence enterprise engineering proactive, personalized intelligence paired with next-generation hardware. The organization moves past standard chat interfaces to build agentic systems that seamlessly interact with humans and the physical world through speech, text, vision, and memory.

How to Apply

Interested candidates should visit the official Hark careers page to submit their application for this position.

Questions candidates ask

Where is this job located?

This position is based in San Jose, United States.

What is the salary range for this role?

The annual US base salary range is between $180,000 and $450,000, varying based on factors like experience and skills.

What department is hiring for this position?

The AI Foundation Models department, specifically the Omni team, is hiring for this role.

What core technical skills are required?

Candidates need experience in multimodal or vision-language models, post-training methods like SFT and preference optimization, video modeling, and data-driven experimentation.

How can I apply to Hark?

You can apply by visiting the official Hark careers page.

Ready to apply?

The application is on the official Hark careers page.

Apply on company site