Research Engineer

Taste Labs; San Francisco (CA); Presencial; USD 140.000 - 220.000; Tempo integral

Resumo da vaga

Taste Labs in San Francisco is seeking a versatile ML engineer to own end-to-end experimentation pipelines—from training runs to evaluation and post-training infra. You’ll publish results and collaborate with AI labs on frontier projects.

You’ll build reward models, judges and verification tools for subjective tasks like design, writing, and visual style; you’ll run post-training experiments on open-source models and shape Taste’s evaluator stack.

Qualificações

• Experience with LLMs and evaluation of subjective domains.
• Ability to design and run post-training experiments.
• Experience building end-to-end ML pipelines and evaluators.
• Familiarity with research-style thinking in a startup setting.

Responsabilidades

• Train reward models, classifiers, judges, and verifiers for subjective domains (e.g. design, writing, visual style).
• Develop frontier evaluations and benchmarks for subjective domains.
• Run post-training experiments on open-source models to test new data formats and post-training techniques.
• Setup the infrastructure to run this.
• Collaborate with AI labs and creative experts to design pilots and experiments around taste.
• Own the end to end pipeline.
• Publish blogs and whitepapers.

Conhecimentos

LLM experience

Post-training

Eval/Judges

Full-stack ML

Infra setup

Research mindset

Creative flair

Ferramentas

Python

PyTorch

TensorFlow

Open-source models

Descrição da vaga

About the company
Taste Labs is building the data and infrastructure layer for taste.
Our goal is to end AI slop. To make AI feel right, not just be correct. We raised $18.5M in seed co-led by Amplify and CRV, and most frontier labs are already customers.
AI has nailed objective domains and can generate anything. The hard part left is judgement: what fits, what feels like you, what's actually GREAT. We're turning that into something measurable, starting with design.
We do it on two sides: building the post-training data and RL environments that teach taste to frontier models, and the context and verification tools agents need to produce work that's more creative, more on-brand, more right.
If that problem excites you, you'll like it here!

About the role
You’ll own full-stack ML experiments, training runs, infra, data and eval pipelines, and publishing results. You’ll develop Taste’s internal research, evaluation design, judges and reward models while collaborating directly with AI labs on frontier projects.

What You'll Do

• Train reward models, classifiers, judges, and verifiers for subjective domains (e.g. design, writing, visual style).
• Develop frontier evaluations and benchmarks for subjective domains.
• Run post-training experiments on open-source models to test new data formats and post-training techniques.
• Setup the infrastructure to run this.
• Collaborate with AI labs and creative experts to design pilots and experiments around taste.
• Own the end to end pipeline.
• Publish blogs and whitepapers.

You Might Be a Good Fit If You

• Are obsessed with taste and want a world with less AI slop.
• Experience in post-training and in evals/judges
• LLM experience required, cannot be purely classical ML
• Thinks like a researcher, moves like an engineer. Creative, scrappy, comfortable in ambiguity
• Startup experience strongly preferred; deep passion for creativity and design