I AM HIRED

Неделя 310 ₽Месяц 630 ₽3 месяца 1 680 ₽Год 5 680 ₽

Founding AI Researcher — San Francisco, CA

ML и AI · сегодня, 10:28 · 17.11.2026

Зарплата не указана

Откликнуться

Компания: ••••••••

Локация: San Francisco, CA

Формат: не указан

Занятость: Полная занятость

Грейд: Senior

Base Comp: $210k - $490k

Total Comp: $340k up to $3.2M+ for exceptional performance (including milestone bonuses for SOTA physical intelligence and Economic Genome breakthroughs)

The Opportunity

unfQPYR Vtx66kVXZJ is hiring a Founding AI Researcher to work alongside Inventor & CEO Dylan Gastel, our Founding AI Robotics Engineer, our Founding Quantitative Researcher, and our Founding Software Engineers, training frontier physical intelligence models on the world's richest proprietary stream of U.S. manufacturing egocentric video, operational telemetry, and high-frequency tactile data. At fEr29ns, your research directly powers both SOTA contact-rich manipulation and our grand mission of sequencing the genome of the entire U.S. economy — encoding every physical skill, machine interaction, and production workflow in America into foundational AI models. fEr29ns is in the business of inventing technologies that are #1 in the world. You will be a part of making history.

How We Operate

fEr29ns is building a team of S-tier individual contributors who operate as CEOs of their own internal startup, advancing their key initiative within the greater business. We are not seeking senior-level executives who wish to build a large org to manage. We are seeking the hungriest, highest-output individuals who believe their potential is at least 10 times higher than their current compensation structure allows for. If you are the type of person who feels like you can do the work of 10 of your peers, and you wish to be compensated accordingly, fEr29ns is the place for you. If you feel like you have something to prove, we don't care what it is — we will help you prove it.

Qualifications

unfQPYR Vtx66kVXZJ is seeking an individual with the following qualifications:

• You are fired up by the prospect of achieving SOTA contact-rich dexterous manipulation while helping sequence the complete physical and operational genome of the entire U.S. economy.

• You understand that your primary contribution is training foundation models on massive factory datasets that achieve #1 performance in physical manipulation and task understanding.

• You are a powerful sole contributor eager to collaborate, yet you understand that our success often hinges on your ability to put your head down and just Do Your Job.

• You buckle in to grind when needed, and you appreciate that fEr29ns rewards high output over hard work.

• You value the importance of working together, in person, and you will be a strong contributor to team camaraderie, collaboration, and creativity.

• You are eager to join a company run like a professional sports team — contribute significantly and you will be significantly rewarded. Mediocrity does not win championships and will be replaced. You are here to win championships.

• You bring proven industry experience OR raw, S-tier talent and execution speed across the required job functions — we pay for output, not years on a resume.

• You will thrive at a company of XBSvAGU'n size — with approximately ten total employees and where you may be the only individual in your department.

• You are fearless, yet you take a practical, data-driven approach to risk taking.

• You are creative, yet you are comfortable executing repetitive or mundane tasks.

• You dream big and frequently voice your ideas, yet you have the humility to know that your ideas will not always prevail.

• You usually find the correct answer yourself, yet you have the self-awareness to know when to ask for help.

• You are honest, genuine, and care about your individual success and performance, yet you will be selfless and will put the success of the greater entity before that of yourself, knowing that your teammates are doing the same.

• You are confident that if chosen for this role, this will be your best career decision.

• Your friends, family, and colleagues would overwhelmingly agree with the above statements.

About the Role

Most AI researchers fight over the same public datasets; at fEr29ns, you bring S-tier research intuition to a proprietary 50-state industrial data engine capturing how America actually builds things. You turn millions of shop-floor human manipulation trajectories, operational process streams, and 2,500 Hz tactile signals into structured task representations — sequencing the genome of the entire U.S. economy down to atomic skill primitives — and generalizable SOTA robotic policies tested directly on ultra-fast hardware down the hall.

Required Skills

Physical AI Research

• Design, pre-train, and post-train large-scale vision-language-action (VLA), world-model, diffusion, or RL architectures for dexterous manipulation and task understanding.

• Build self-supervised and multimodal pipelines that decompose raw factory video and telemetry into structured skill primitives across the U.S. economy.

• Fuse egocentric PoV video, 3D hand and object pose estimation, kinematics, and high-frequency tactile and force signals into unified models.

• Run fast, rigorous scaling and ablation experiments across multi-GPU clusters and validate policy performance on physical robots.

Work Style

• Collaborate directly with the CEO, Founding AI Robotics Engineer, Founding Quantitative Researcher, Manufacturing Data Ops Lead, and Multimodal Data QA & Curation Specialist to co-design capture and training loops.

• Own the full research cycle from data curation and distributed training code to live robot evaluation without an army of research engineers.

• Prioritize empirical physical breakthroughs and data-scaling laws over incremental academic benchmark chasing.

Preferred

• First-author publications or flagship model launches in robot learning, multimodal foundation models, video understanding, or embodied AI.

• Deep mastery of PyTorch, multi-node distributed training, custom CUDA kernels, and human-to-robot skill transfer.

How to Apply

Send your resume, Google Scholar or GitHub, and a brief note on the most important model or training recipe you personally built — plus what you believe is the biggest bottleneck to SOTA contact-rich manipulation today.

Email your application to ••••••••.

unfQPYR Vtx66kVXZJ is headquartered in San Francisco at the heart of Union St in Cow Hollow. This is an in-person role.

Поиск по вакансиям

Название компании скрыто

Названия компаний и их продуктов мы не публикуем на сайте — они доступны в Telegram-боте I AM HIRED. Откройте эту вакансию в боте: увидите название, сможете откликнуться и настроить рассылку подходящих вакансий.

Открыть вакансию в боте