Research Engineer - Post-Training

Job Title

Research Engineer — Post-Training

Role Summary

Post-train frontier models to autonomously perform complex tasks across the semiconductor design and verification pipeline. Work with hardware design and verification experts to build reinforcement learning environments, reward functions, and evaluation and benchmarking frameworks that improve model reliability, efficiency, and creativity in semiconductor reasoning.

Experience Level

Mid-level. The posting does not specify years of experience; candidates should have practical experience training or fine-tuning large models and building RL-based training/evaluation pipelines.

Responsibilities

Primary responsibilities include designing environments, training pipelines, and evaluations that enable models to propose, optimize, and verify chip designs.

• Design and implement RL or curriculum-learning environments for LLMs or multimodal agents that model chip design workflows.
• Develop structured reward functions and feedback pipelines balancing correctness, performance, and design efficiency.
• Create and scale high-quality evaluation datasets and benchmarks for complex reasoning and design tasks.
• Run large-scale RL fine-tuning and post-training experiments; manage training experiments and infrastructure.
• Collaborate with domain experts in hardware design and verification to define metrics, constraints, and simulation conditions.
• Analyze model failures, identify verification gaps, and iterate on environments, rewards, and evaluation criteria.

Requirements

Core qualifications and desirable skills.

• Must-have: Practical experience creating and scaling RL environments for LLMs or multimodal agents.
• Must-have: Experience running large-scale RL fine-tuning or post-training experiments for frontier models.
• Must-have: Ability to design reward functions and feedback pipelines for structured reasoning or symbolic domains.
• Must-have: Experience building evaluation datasets and benchmarks for complex reasoning or design tasks.
• Must-have: Experience collaborating with domain experts to incorporate constraints and define evaluation metrics.

Nice-to-have:

• Experience applying reinforcement learning or curriculum learning to hardware design, verification, or symbolic reasoning domains.
• Familiarity with semiconductor design workflows, RTL, simulation, or verification tools.

Education Requirements

Not specified.

About the Company

Company: Voltai

Headquarters: Palo Alto, CA, United States

Voltai develops AI-driven world models and agents to design, evaluate, and optimize physical systems—focusing on hardware, electronics, and semiconductors to enable AI-led hardware co-design, performance modeling, and cross-domain optimization.