Machine Learning Engineer – Data for Foundational Models

About the role

You will own the data infrastructure that powers Sarvam's next family of foundational models, building petabyte‑scale curation and filtering pipelines and ensuring data quality with the rigor of a research team.

Key responsibilities

• Design and build large‑scale data pipelines for pre‑training and post‑training at petabyte scale, covering ingestion, parsing, normalization, filtering, deduplication, tokenization, and packing.
• Develop and continuously improve quality‑filtering systems, including model‑based classifiers and contamination detection.
• Own data mixture design, curriculum, and annealing strategies in partnership with the research team.
• Build tooling that enables researchers and engineers to analyze, slice, attribute, and debug data.
• Scale pipelines to handle multilingual corpora, code, math, web data, and licensed datasets while tracking provenance and licensing end‑to‑end.

Required profile

• BS or MS in Computer Science or a related technical field, or equivalent demonstrated experience.
• 3+ years of experience building large‑scale data systems (petabyte‑scale processing, distributed pipelines).
• Hands‑on experience with data curation and filtering for LLM training.
• Deep familiarity with distributed data processing frameworks (Spark, Ray, Beam, Dask) and underlying storage systems.
• Strong Python skills and understanding of low‑level data‑path components such as tokenization, sharding, packing, and I/O patterns.

Required skills

• Python
• Spark
• Ray
• Beam
• Dask