Member of Technical Staff, Data

Inception; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Inception in San Francisco is seeking experienced engineers and scientists to shape how we collect, process, and curate datasets powering our models. You will build scalable data pipelines, develop synthetic data generation techniques, and ensure models train on high-quality, diverse data.

The role combines engineering with research, requiring hands-on implementation with Python, Spark, and ML frameworks. You will design data ingestion systems, manage large-scale storage, and create evaluation

Qualifications

• 3+ years of experience building data processing pipelines at scale for AI/ML.
• Strong proficiency in Python and experience with data processing frameworks (Spark, Beam, Airflow).
• Familiarity with synthetic data generation techniques and data augmentation strategies.
• Familiarity with web scraping, crawling technologies, and Common Crawl datasets.
• Solid understanding of machine learning fundamentals and experience with ML frameworks (PyTorch, TensorFlow).
• Experience with SQL and NoSQL databases for managing structured and unstructured data.

Responsibilities

• Develop data mixes for training LLMs, including leveraging open-source datasets, synthetically generated data, and curated human feedback.
• Design and implement data pipelines for processing petabyte-scale datasets.
• Build systems for web crawling, data ingestion, and real-time data processing to support model training.
• Develop tools and frameworks for efficient data storage, retrieval, and versioning across distributed systems.
• Create evaluation frameworks to measure data diversity, quality, and representativeness.
• Ensure data collection adheres to privacy regulations.

Skills

Python

Data pipelines

ML fundamentals

Education

BS/MS/PhD in CS/ML or related field

Tools

Apache Spark

Beam

Airflow

PyTorch

TensorFlow

SQL

NoSQL

Common Crawl

S3

BigQuery

Job description

The Role

We seek experienced engineers and scientists to shape how we collect, process, and curate the datasets that power our models. You\'ll combine engineering expertise with research insight to build scalable data pipelines, develop synthetic data generation techniques, and ensure our models are trained on high-quality, diverse data.

Key Responsibilities

• Develop data mixes for training LLMs, including by leveraging open-source datasets, synthetically generated data, and curated human feedback.
• Design and implement data pipelines for processing petabyte-scale datasets.
• Build systems for web crawling, data ingestion, and real-time data processing to support model training.
• Develop tools and frameworks for efficient data storage, retrieval, and versioning across distributed systems.
• Create evaluation frameworks to measure data diversity, quality, and representativeness.
• Ensure data collection adheres to privacy regulations.

Qualifications

• BS/MS/PhD in Computer Science, Machine Learning, or a related field (or equivalent experience).
• 3+ years of experience building data processing pipelines at scale, particularly with AI/ML applications.
• Strong proficiency in Python and experience with data processing frameworks (Apache Spark, Beam, Airflow).
• Familiarity with synthetic data generation techniques and data augmentation strategies.
• Familiarity with web scraping, crawling technologies, and Common Crawl datasets.
• Solid understanding of machine learning fundamentals and experience with ML frameworks (PyTorch, TensorFlow).
• Experience with SQL and NoSQL databases for managing structured and unstructured data.

Preferred Skills

• Experience with large language models and understanding of tokenization, embeddings, and model architectures.
• Experience managing human annotation workflows and quality control processes.
• Experience with vector databases and embedding-based retrieval systems.
• Knowledge of data privacy regulations and ethical AI practices.
• Experience with distributed computing and large-scale data storage systems (HDFS, S3, BigQuery).