Open to opportunities

SIDDHARTHA NAG

@siddharthanag

Message

LLM and GenAI data specialist training and validating high-quality multimodal datasets.

India

Message

What I'm looking for

I’m looking for work where I can run reliable LLM/GenAI data annotation pipelines, do QA and validation, and collaborate with ML teams to tighten labeling standards and improve throughput.

I curate and annotate high-quality training datasets for LLM and GenAI systems, including multimodal work across text, image, and structured data, and I validate annotations through peer review and QA workflows. At Innodata, I’ve supported fine-tuning and RLHF-style processes while improving annotation guideline consistency.

Earlier, at Telus Digital, I managed large-scale ML and LLM labeling pipelines with 98%+ accuracy and partnered with AI/ML engineers to refine standards based on labeling schemas and ambiguity findings. Before that, I handled geocoder QA at Amazon Development Centre, where I verified millions of coordinates, built SQL/Excel QA dashboards, and reduced processing errors by implementing corrective actions and standardized protocols.

Experience

Work history, roles, and key accomplishments

Data Annotator

Innodata

Jan 2026 - Apr 2026 (3 months)

Curated and annotated high-quality training datasets for LLM and GenAI systems, ensuring compliance with project-specific annotation schemas and quality benchmarks.

Data Annotation LLM GenAI Quality Assurance Peer Review

Data Annotator

Telus Digital

Sep 2025 - Jan 2026 (4 months)

Managed large-scale AI training datasets for machine learning and LLM pipelines, ensuring high-quality labeled data with 98%+ accuracy.

Data Annotation AI Training Data Machine Learning LLM Quality Assurance

Quality Specialist (Geocoder)

Amazon Development Centre

Mar 2022 - Aug 2022 (5 months)

Ensured 99%+ accuracy in geospatial data by verifying and correcting geographic coordinates for millions of addresses across Amazon's mapping systems.

Geospatial Data QA Testing SQL Excel Data Validation geocoding