Research Scientist (Frontier AI Infra)

Greylock Partners; New York (NY); On-site; USD 180,000 - 240,000; Full time

Job summary

Greylock Partners is seeking researchers for a frontier AI role that blends AI research with software engineering and real-world expertise. You’ll work across model evaluation, benchmark design, reasoning, data generation, applied AI systems, and research infrastructure.

Some weeks involve creating new evaluation methods, others translate research into production-ready systems with measurable impact. You’ll collaborate with founders and the early team across research, engineering, product, and

Qualifications

• Experience with large language models or modern AI systems
• Model evaluation, benchmarking, reasoning, or reinforcement learning
• Publishing research at top-tier conferences or equivalent impact
• Building production-quality research software in Python and modern ML frameworks
• Translating research ideas into practical systems with measurable impact
• Operating effectively in ambiguous, fast-moving startup environments

Responsibilities

• Designing novel evaluation frameworks for frontier AI models
• Developing benchmarks that measure increasingly sophisticated model capabilities
• Conducting original research on model improvement through data, evaluation, and applied AI systems
• Building research tooling and infrastructure that accelerates experimentation
• Translating complex real-world workflows into rigorous evaluation and model improvement programs
• Working closely with engineering and domain experts to rapidly move research into production
• Helping shape the technical direction of an early frontier AI infrastructure company

Skills

LLMs

Model evaluation

Benchmarking

Python

ML frameworks

Research translation

Tools

PyTorch

TensorFlow

Job description

Seed-stage frontier AI infrastructure company with offices in San Francisco and New York building the next generation of software, data, and evaluation systems for frontier AI.

The first wave of frontier AI was driven by scaling compute and model architecture. As foundation models become increasingly capable, one of the next major bottlenecks is helping them reason more effectively, master complex real-world work, and continuously improve across specialized domains.

This company is building for that shift.

Working directly with leading frontier AI organizations, the team develops novel evaluation frameworks, applied AI systems, and real-world data that help advance the capabilities of frontier models. The founding team brings deep experience from Scale AI, Meta AI, and multiple successful startups, and is assembling an exceptionally technical, in-person team spanning research, engineering, and operations.

Summary

This isn't a traditional AI research role.

The company is looking for researchers who enjoy tackling difficult, open-ended problems that sit at the intersection of AI research, software engineering, and real-world expertise.

You’ll work across model evaluation, benchmark design, reasoning, data generation, applied AI systems, and research infrastructure. Some weeks you’ll be developing entirely new ways to measure model capabilities. Others you’ll partner with engineers and domain experts to translate complex real-world workflows into rigorous evaluation and improvement systems.

If you're excited by difficult technical problems, high ownership, and helping shape the next generation of frontier AI capabilities, this is one of the more compelling opportunities we've seen.

What You’ll Own

• Designing novel evaluation frameworks for frontier AI models
• Developing benchmarks that measure increasingly sophisticated model capabilities
• Conducting original research on model improvement through data, evaluation, and applied AI systems
• Building research tooling and infrastructure that accelerates experimentation
• Translating complex real-world workflows into rigorous evaluation and model improvement programs
• Working closely with engineering and domain experts to rapidly move research into production
• Helping shape the technical direction of an early frontier AI infrastructure company

You’ll work directly with the founders and early team across research, engineering, product, and operations.

What We’re Looking For

You likely have experience with several of the following:

• Research in large language models, foundation models, or modern AI systems
• Model evaluation, benchmarking, reasoning, post-training, reinforcement learning, synthetic data, or AI agents
• Publishing research at top-tier conferences or demonstrating equivalent research impact
• Building production-quality research software in Python and modern ML frameworks
• Translating research ideas into practical systems with measurable impact
• Operating effectively in ambiguous, fast-moving startup environments

Experience at frontier AI labs or leading AI research organizations is highly relevant, but we’re primarily looking for exceptional researchers who enjoy solving difficult technical problems and rapidly turning research into real-world impact.

Prior startup experience is a strong plus.

Our ideal candidate is intellectually curious, builder-oriented, highly collaborative, and excited to help define the next generation of frontier AI infrastructure.

About Us

Greylock is an early-stage investor in hundreds of remarkable companies including Airbnb, LinkedIn, Dropbox, Workday, Cloudera, Facebook, Instagram, Roblox, Coinbase, Palo Alto Networks, among others. Learn more at https://greylock.com/

How We Work

We are full-time, salaried employees of Greylock and provide free candidate referrals and introductions to our active investments. This posting is for direct employment with one of our portfolio companies. We review every application and reach out directly when we believe there’s a strong potential fit.