Data & Annotation Engineer: CVAT Ops & Provenance

Innodata Inc.; Washington (District of Columbia); On-site; USD 75,768 - 82,656; Full time

Job summary

Innodata invites you to join its Federal Practice as a Data/Annotation Engineer, focusing on building a governed data layer for sensitive datasets. You will administer the CVAT toolchain, manage annotation workflows across the corpus, and ensure data is accurate, well-labeled, and fully traceable for model training.

You’ll work with the AI Solutions Engineer, produce per-dataset documentation for governance, and drive high-quality data through Phase D corpus production to support critical Trust

Qualifications

• Bachelor's degree in Data Science, CS, or related field (or equivalent experience).
• 5+ years total professional experience; 3+ years in data engineering or annotation operations.
• CVAT deployment and day-to-day operation required.
• Annotated dataset ingest pipelines: schema mapping, format validation, ontology alignment.
• Full-motion video (FMV) annotation concepts and tooling.
• Python scripting for data wrangling, validation, and format conversion.
• Active Secret clearance with TS/SCI eligibility.

Responsibilities

• Receive, validate, ingest, and ontology-map the ODIN corpus from AFS delivery
• Produce the ODIN load report: corpus description, ontology mapping, readiness state
• Configure CVAT annotation pipeline against the Phase 1 starter kit rule pack
• Operate both self-service and lightweight white-glove annotation paths during Phase D corpus production
• Produce 50-100 label demonstration corpus across synthetic and mission-aligned content
• Support QA/Evaluation Lead on QC execution and corpus annotation dry-runs
• Associate DataCard provenance records with annotated and synthetic outputs in coordination with the Solution Architect

Skills

Python scripting

Annotation operations

FMV annotation concepts

Ontology mapping

Security clearance

Education

Bachelor's degree in Data Science / Computer Science

Tools

CVAT annotation platform

Job description

Innodata invites you to join its Federal Practice as a Data/Annotation Engineer, focusing on building a governed data layer for sensitive datasets. You will administer the CVAT toolchain, manage annotation workflows across the corpus, and ensure data is accurate, well-labeled, and fully traceable for model training.

You’ll work with the AI Solutions Engineer, produce per-dataset documentation for governance, and drive high-quality data through Phase D corpus production to support critical Trust