Data Infrastructure Engineer (AI)

Location: On-site in San Francisco, California; the listing also mentions New York.

Annual salary: $300,000–$400,000 USD, depending on background, skills, and experience.

Company

Thinking Machines Lab builds AI systems designed to extend human will and judgment, including frontier models, model customization tools, and human-AI interfaces.

What you will do

• Design, build, and operate fault-tolerant infrastructure for LLM research, including distributed compute, data orchestration, and multimodal storage.
• Develop high-throughput systems for data ingestion, processing, transformation, training data catalogs, deduplication, quality checks, and search.
• Build traceability, reproducibility, and quality-control systems across the data lifecycle.
• Implement monitoring and alerting for platform reliability and performance.
• Collaborate with research teams to improve data quality, accelerate experiments, and shorten training cycles.

Requirements

• Bachelor’s degree or equivalent experience in computer science, engineering, or a related field.
• Proficiency in at least one backend language: Python or Rust.
• Fluency with distributed compute frameworks such as Apache Spark or Ray.
• Strong familiarity with cloud infrastructure, data lake architectures, and batch and streaming pipelines.
• Experience with Kafka, dbt, Terraform, Airflow, web crawlers, deduplication, data mining, search, file formats, and storage systems such as Parquet and Delta Lake.
• Ability to own projects end-to-end, work across technology stacks, and collaborate with research and cross-functional teams.

Culture & Benefits

• Small, high-impact engineering team working closely with researchers and subject matter experts.
• Generous health, dental, and vision benefits.
• Unlimited paid time off and paid parental leave.
• Documentation, testing, and teammate-enablement are emphasized.
• Visa sponsorship is available, with support through the visa process for qualified candidates.
• Relocation support is available as needed.