Research Engineer, Web-Scale Content Understanding

Speedrun Talent Network; Zürich; Hybrid; CHF 140,000 - 190,000; Full time

Job summary

Exa is an applied AI lab building a groundbreaking search architecture. As a backend engineer, you’ll contribute to parsing, classification, and extraction at web scale, choosing projects aligned with your strengths.

We value graduate-level ML experience, the ability to build transformers in PyTorch, and a passion for solving hard problems where ground truth is evolving. You’ll work on credible model design and scalable data pipelines.

Qualifications

• Graduate-level ML experience (Master’s or PhD with at least 2 years of relevant experience), or an exceptionally strong undergrad.
• Ability to build a transformer from scratch in PyTorch, and run models cheaply at scale.
• Love for building large-scale datasets; focus on supervision rather than architecture.
• Comfort with problems where ground truth does not exist yet and defining it is part of the job.
• Care about finding high-quality knowledge with broad-world impact.

Responsibilities

• Explore parsing and classification of pages to improve retrieval quality.
• Develop supervision strategies for quality judgments and model alignment.
• Investigate credibility and misinformation as modelling problems at web scale.
• Assess semantic similarity of documents to aid deduplication without losing useful pages.
• Design classification and extraction components that scale affordably.
• Lead experiments to define labels in areas lacking existing ground truth.

Skills

Graduate-level ML experience

Transformer models in PyTorch

Large-scale dataset creation

Ground-truth design & evaluation

Education

Master’s or PhD in ML/CS

Tools

PyTorch

Job description

Exa is an applied AI lab building a groundbreaking search architecture. As a backend engineer, you’ll contribute to parsing, classification, and extraction at web scale, choosing projects aligned with your strengths.

We value graduate-level ML experience, the ability to build transformers in PyTorch, and a passion for solving hard problems where ground truth is evolving. You’ll work on credible model design and scalable data pipelines.