RC RANDOM CHAOS

Jane Street Redesigns Coding Interviews for the LLM Era

· via Hacker News

Original source

Our software engineering interviews are changing a little

Hacker News →

Jane Street acknowledges that machine learning models have fundamentally altered development workflows across their organization—most engineers now integrate them into code writing, reading, and debugging routines. This widespread adoption prompted the firm to reconsider how it evaluates engineering candidates, recognizing that their interview approach no longer matches how the job has actually evolved.

Starting this summer, the firm is piloting revised interview formats for campus hires. The new approach emphasizes software design, code quality assessment, and strategic use of AI tools while reducing emphasis on raw code-writing ability. Most interview rounds will remain traditional coding assessments conducted without LLM assistance, but some will test higher-level decision-making. The exact format is still in flux as the firm experiments, so candidates will receive advance notice of what to expect.

Jane Street clarifies that deep LLM expertise isn’t a prerequisite for success. The firm values how candidates reason about tool selection—understanding what each tool excels at and knowing when to use alternatives. As code generation becomes cheaper, other capabilities have grown more valuable: choosing what to build, selecting lasting abstractions, and verifying correctness. The interview changes reflect a shifted focus toward evaluating these judgments.

Read the full article

Continue reading at Hacker News →

This is an AI-generated summary. Read the original for the full story.