Staff Parser/Compiler Research Engineer

dottxt · Remote · Remote - USA ·
Fully remoteFull-timeSenior5+ years
COMPENSATION
$160k – $210k
0 applicants so far

Listed on the employer's job board (Ashby)

  • Staff level role
  • Fully remote (US only)
  • Python + Rust/C++
  • Parser research focus
  • Stock options included

You will work on the core technology that allows language models to produce reliable, formatted output. By compiling schemas and grammars into token constraints, you enable safe and deterministic agent behavior.

The team focuses on integrating advanced parser research with LLM decoding loops. You will design efficient data structures and ensure the system runs fast enough for real-time use.

What you'll do

  • Design high-performance parsers and data structures for LLM integration
  • Research advanced compiler technology and symbolic methods
  • Optimize code for reliability and efficiency within decoding loops
  • Implement rigorous testing and automation for core methods
  • Stay updated on relevant advancements in large language models

What you'll need

  • Professional proficiency in Python required
  • Practical experience with lower-level languages like C, C++, or Rust
  • Deep knowledge of parsers such as Earley or LALR(1)
  • Experience with formal grammars and finite-state machines
  • Ability to implement efficient data structures from scratch
  • Strong written communication skills for concise documentation
  • Familiarity with .txt technical reports and open-source projects

Stack

  • Python
  • C
  • C++
  • Rust

How you'll work

  • Fully remote company structure
  • Must be currently based in the United States
  • Communication is primarily written
  • Occasional organic discussions replace large meetings

What you get

  • Market-rate salary compensation
  • Equity stock options provided
  • Health and dental insurance coverage
  • 401(k) plan for US employees
  • Company-provided GPU hardware

Hiring process

  • Submit a one-page resume in English
  • Apply to only one position at a time