Role Overview
The Member of Technical Staff, Post-Training at Handshake is a high-ownership position focused on designing and building advanced post-training methodologies for frontier AI models. Based in San Francisco, California, with remote flexibility, this full-time role suits builder-researchers who want to transform complex AI challenges into scalable production systems.
Key Responsibilities
Create and execute post-training strategies for frontier models, incorporating techniques like supervised fine-tuning, reinforcement learning, preference optimization, and reward modeling.
Convert open-ended research questions and partner needs into actionable hypotheses, rigorous experiments, and robust production code.
Construct and refine evaluation frameworks, training environments, performance benchmarks, and data-processing pipelines.
Execute rapid iteration loops by prototyping solutions, assessing performance, analyzing outcomes, and scaling successful learnings into durable products.
Collaborate closely with domain experts and AI researchers to establish high-signal data systems and evaluation metrics.
Generalize repeatable patterns from partner engagements into scalable, reusable software platforms.
Elevate engineering standards through clean code, strong architectural judgment, and technical mentorship.
Advance the broader AI field by contributing to open-source tools, technical publications, and industry benchmarks.
Requirements
Minimum of three years of proven experience in model evaluation, fine-tuning, or post-training workflows, such as SFT, RLHF, DPO, PPO, LoRA/PEFT, or reward modeling.
Advanced proficiency in Python, with the ability to develop clean, scalable software systems.
Practical background utilizing modern machine learning frameworks, specifically PyTorch alongside large-scale data and evaluation pipelines.
Exceptional experimental rigor, including the capability to formulate hypotheses, isolate critical metrics, diagnose failures, and interpret data.
Demonstrated system design expertise capable of balancing tradeoffs concerning scale, quality, reliability, and reuse.
Ability to thrive with substantial ownership inside an ambiguous and fast-paced operational setting.
Low-ego communication style that fosters effective teamwork across multidisciplinary groups.
Nice to Have
Background in operating or constructing large-scale machine learning training, inference, data, or evaluation architectures.
Experience designing agent benchmarks, LLM evaluation processes, annotation systems, or data-quality verification frameworks.
Applied research history centered on reinforcement learning, model alignment, synthetic data generation, or human-in-the-loop platforms.
A public track record of published research, open-source software contributions, or technical leadership within machine learning.
Demonstrated success productizing custom research into standardized, reusable software platforms.
Skills and Tools
Python
PyTorch
Large Language Models (LLMs)
Supervised Fine-Tuning (SFT)
Reinforcement Learning from Human Feedback (RLHF)
Direct Preference Optimization (DPO)
Reward Modeling
Salary and Benefits
Annual salary range between $200,000 and $350,000 USD.
Company equity providing ownership in a rapidly expanding business.
Comprehensive financial wellness options including a 401(k) match, financial coaching, and competitive pay.
Family support benefits featuring paid parental leave, fertility support, and parental coaching.
Extensive health coverage including medical, dental, vision, mental health assistance, and a $500 wellness stipend.
Professional development resources including a $2,000 learning stipend and ongoing career coaching.
Office perks in San Francisco featuring free daily lunch, commuting support, and an on-site gym.
Flexible paid time off, 15 designated holidays, plus two flexible days.
Team connection activities and referral bonuses.
About the Company
Handshake initially built a career platform connecting millions of job seekers, employers, and educational institutions. In 2025, the organization launched Handshake AI to collaborate directly with frontier AI lab researchers, publish industry benchmarks, and develop innovative external AI products through Handshake Labs.
How to Apply
Please visit the official Handshake careers page to submit your application for this position.
