Tax Form Annotation Specialist

About Company:

Our client is a Palo Alto–based AI infrastructure and talent platform founded in 2018. It helps companies connect with remote software developers using AI-powered vetting and matching technology. Originally branded as the “Intelligent Talent Cloud, "enabled companies to “spin up their engineering dream team in the cloud” by sourcing and managing vetted global talent.

In recent years, they have evolved to support AI infrastructure and AGI workflows, offering services in model training, fine-tuning, and deployment—powered by their internal AI platform, ALAN, and backed by a vast talent network. They reported $300 million in revenue and reached profitability. Their growth is driven by demand for annotated training data from AI labs, including major clients like OpenAI, Google, Anthropic, and Meta.

Job Description:

Job Title: Tax Form Annotation Specialist

Location: Pan India

Experience: 4+ yrs.

Employment Type: Contract to hire

Work Mode: Remote

Notice Period: - Immediate joiners

Roles & Responsibilities:

Role Overview

We are seeking an efficient and detail-oriented Tax Form Annotation Specialist. This hybrid role requires both hands-on experience with tax documentation and the technical expertise to annotate forms in Label Studio or any other software. You will use tax forms to extract bounding box information with correct relationships between the boxes.

Core Workflow & Responsibilities

The role follows a structured six-step pipeline:

Phase 1 — Research

• Form Research: Study tax form layouts, field relationships, and IRS/Financial institutions/local authority instructions to establish annotation scope and rules.

Phase 2 — Drawing Bounding Boxes

• Define Field Keys & Schema: Create Labels, Fields, and Data Types in Label Studio for every identifiable field on the form.
• Draw Bounding Boxes: Precisely draw bounding boxes over each field and assign the correct Label, Field name, and Data Type to every box.

Phase 3 — Relationships

• Define Inter-field Edges: Identify intercorrelated fields and configure relationship tags and edge links between them in Label Studio.
• For calculation relations, provide formula in a separate sheet to be used in Phase 4

Phase 4 — Post-Processing

• Export, Test & Validate: Generate JSON and YAML annotation exports, test overlay alignment on source PDFs, and verify schema integrity.

Required Qualifications & Skills

Tax & Domain Knowledge

• Solid understanding of tax form structures (e.g. W-2, 1040, 1099, VAT returns, Self-Assessment, corporate tax schedules)
• Familiarity with field dependencies within tax forms — relationships between income lines, deductions, credits, and totals
• Working knowledge of tax authority form-filling guidelines (IRS, Financial institutions, or equivalent)
• Ability to research and interpret unfamiliar form types independently

Annotation & Technical Skills

• Hands-on experience with Label Studio or a comparable document annotation platform
• Proficiency in drawing precise bounding boxes and applying structured label schemas
• Understanding of field-level data types: strings, dates, currency, checkboxes, enumerations
• Ability to define and document field metadata including format constraints and validation rules
• Familiarity with JSON and YAML for reviewing and validating annotation exports
• Experience linking related fields using relationship/edge tagging

Nice to Have

• Background in accounting, bookkeeping, or tax preparation
• Experience with bounding boxes and label studio or similar
• Prior work on training data for document AI

What Success Looks Like

• Bounding boxes are accurate, consistent, and cover all meaningful fields on every form
• Field keys and schemas are correctly defined and follow agreed naming conventions
• Metadata is complete and correct for every annotated field
• Exported JSON/YAML overlays cleanly on source PDFs with no misalignment
• Inter-field relationships are logically mapped and documented
• Adapting to changing requirements in a fast paced work environment