Member of Technical Staff, Model Training (post-training research)

Zekai; United States; Remote; USD 200,000 - 320,000; Full time

Benefits offered by this job

Equity compensation

Bonuses

Remote-first workforce

Health insurance reimbursement

401(k) with company match

Job summary

Zekai is seeking a Member of Technical Staff, Model Training, focusing on post-training research. This is a remote role that blends fundamental ML research with large-scale model development.

You will design and run ambitious experiments, translate research into scalable training methods such as SFT, RL, and distillation, and contribute to model intelligence. The position offers a remote-first environment with strong compensation and growth opportunities.

Qualifications

• Strong foundations in machine learning, deep learning, optimization, and modern model training techniques.
• Hands-on experience training, fine-tuning, or post-training large-scale language or multimodal models.
• Deep technical expertise in one or more areas such as reinforcement learning, model distillation, reward modeling, data selection, or learning dynamics.
• Strong proficiency in Python and PyTorch, with the ability to implement, debug, and optimize complex training pipelines.
• Experience designing controlled experiments, interpreting model behavior, and translating empirical findings into actionable research directions.
• Ability to independently navigate ambiguous research problems, formulate hypotheses, and iterate quickly based on experimental evidence.
• Strong analytical thinking and a research-driven mindset, with an emphasis on reproducibility, scientific rigor, and measurable improvements.

Responsibilities

• Research, implement, and scale post-training methods across large language and multimodal models (SFT, RL, OPD).
• Investigate learning objectives, optimization dynamics, exploration strategies, and credit assignment to improve training stability and efficiency.
• Develop and refine training recipes by optimizing data mixtures, sampling strategies, curricula, and evaluation methodologies.
• Design methods for learning from expert demonstrations, human preferences, corrections, and model-generated trajectories.
• Research co-evolution of generators, reward models, and graders to improve performance and signaling quality.
• Design rigorous experiments, ablation studies, and evaluation frameworks to measure generalization and reasoning.
• Own research initiatives end-to-end from identifying directions to executing large-scale experiments.

Skills

Machine learning fundamentals

Deep learning

Optimization

Experiment design

Reinforcement learning

Research mindset

Model training techniques

Tools

Python

PyTorch

Distributed training

Experiment tracking

Job description

Job Title: Member of Technical Staff, Model Training (post-training research)

Job Type: Full-time

Location: Remote

The Role

In this role, you will investigate how learning algorithms, training data, and evaluation methodologies interact to produce stronger reasoning, generalization, and agent capabilities. You will explore new approaches to reinforcement learning, distillation, and data-efficient training, translating research insights into scalable methods that improve model intelligence.

This is a highly experimental role at the intersection of fundamental machine learning research and large-scale model development. You will have the opportunity to formulate research hypotheses, design ambitious experiments, and develop training methods that directly influence the capabilities of our models.

What You'll Do

• Research, implement, and scale advanced post-training methods, including supervised fine-tuning (SFT), reinforcement learning (RL), and on-policy distillation (OPD), across large language and multimodal models.
• Investigate learning objectives, optimization dynamics, exploration strategies, and credit assignment mechanisms to improve training stability, sample efficiency, and capability acquisition.
• Develop and refine training recipes by jointly optimizing data mixtures, sampling strategies, curricula, and evaluation methodologies.
• Design methods for learning from expert demonstrations, human preferences, corrections, and model-generated trajectories, identifying where human feedback delivers the greatest improvements.
• Research the co-evolution of generators, reward models, and graders, developing approaches that improve both model performance and the quality of the training signals they consume.
• Design rigorous experiments, ablation studies, and evaluation frameworks to measure generalization, reasoning, and performance on realistic downstream tasks.
• Own research initiatives end-to‑end, from identifying promising directions and implementing new algorithms to executing large-scale experiments and validating results.

What We're Looking For

• Strong foundations in machine learning, deep learning, optimization, and modern model training techniques.
• Hands‑on experience training, fine‑tuning, or post‑training large‑scale language or multimodal models.
• Deep technical expertise in one or more areas such as reinforcement learning, model distillation, reward modeling, data selection, or learning dynamics.
• Strong proficiency in Python and PyTorch, with the ability to implement, debug, and optimize complex training pipelines.
• Experience designing controlled experiments, interpreting model behavior, and translating empirical findings into actionable research directions.
• Ability to independently navigate ambiguous research problems, formulate hypotheses, and iterate quickly based on experimental evidence.
• Strong analytical thinking and a research-driven mindset, with an emphasis on reproducibility, scientific rigor, and measurable improvements.

Nice to Have

• Experience with RLHF, RLAIF, preference optimization, or reinforcement learning for reasoning and agentic tasks.
• Familiarity with distributed training, large-scale experimentation, and GPU-based model optimization.
• Experience developing reward models, verifiers, graders, or automated evaluation systems.
• Research contributions to post-training, reinforcement learning, or related areas through publications, open-source projects, or production systems.
• Experience investigating synthetic data generation, curriculum learning, or iterative self-improvement techniques.

Compensation & Benefits Notice

The national pay range for this full-time position is base salary of $200,000 –$320,000. All employees are eligible for equity compensation, and employees may also receive performance‑based bonuses, dependent on role and subject to company policies. our partner provides a comprehensive benefits package, including up to 100% reimbursement for health‑insurance premiums, paid time off, a 401(K) plan with a company match, and additional benefits designed to support a high‑performing, remote‑first workforce.

our partner is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex (including pregnancy, sexual orientation, or gender identity), national origin, age, disability, genetic information, veteran status, or any other characteristic protected by applicable local laws, regulations and ordinances. If you need assistance and/or a reasonable accommodation during the application process, reach out to support@our partner.

Our hiring process utilizes artificial intelligence tools to assist in candidate screening and assessment. Our AI tools are designed to complement, not replace, human decision‑making.

This is a 100% remote contract role, open to applicants worldwide in the countries the employer accepts.