Data Engineer — ML Training Data Pipeline
Summary
Builds and maintains AWS data pipelines that transform raw production traces into high-quality training datasets for LLM fine-tuning — handling ingestion, deduplication, format conversion, quality filtering, and train/test splitting at scale. Core stack is Python (pandas, pyarrow), JSONL processing, HuggingFace Datasets, and AWS S3/EC2, based in Hyderabad or Pune.
What We Expect:
- Build
end-to-end data pipelines: raw trace
ingestion → dedup → format
conversion → quality gating → training-ready datasets
- Process
large-scale JSONL data on AWS S3 (tens of thousands of traces per batch)
- Convert
between chat-completion formats (e.g., OpenAI → Llama 3.1
tool-calling format)
- Implement
smart deduplication and sampling to balance training distribution
- Design
identity-aware train/test splits that measure true generalization
- Build
data validation gates to detect schema drift and format anomalies
- Create
a continuous pipeline that auto-processes new production traces for
retraining
Requirements
- Experience: 6+
years data engineering focused on ML data pipelines
- Python: Strong — pandas,
pyarrow, JSONL processing at scale
- ML
Data Libraries: HuggingFace Datasets, Arrow-based storage
- Data
Formats: Multi-turn conversation/chat data structures and
tokenizer-specific formatting
- Deduplication: Content
hashing, identity-based grouping strategies
- AWS: S3,
EC2, batch processing workflows
Preferred (Not Required): LLM training data prep
(chat templates, tool-calling schemas); Axolotl or similar dataset formats;
data versioning (DVC, LakeFS); browser-automation trace data or Playwright.
Benefits
- Comprehensive Medical Coverage:
Health insurance of INR 5.0 Lakhs for you and your family (up to 6 members), ensuring complete peace of mind. - Robust Protection Plans:
Group Personal Accident Insurance and Group Term Life Insurance to safeguard you and your loved ones. - Retirement Benefits:
PF and Gratuity provided as per standard government regulations. - Flexible Work Options:
Enjoy hybrid work arrangements & flexible working hours. - Generous Leave Policy:
21 days of annual leave, in addition to 10 company-declared holidays. - Employee Well-being Spaces:
Access to a dedicated break-out area with round-the-clock refreshments for relaxation and rejuvenation.