Medical Coding Auditor-1

Allscripts(India) LLP, ultimately a subsidiary of Altera Digital Health Inc.,[Altera India]; Pune District; On-site; INR 1,200,000 - 2,400,000; Full time

Job summary

Veradigm is seeking experienced medical coders to support our enterprise coding capability. You will focus on ensuring accuracy in AI-generated medical codes, chart preparation, and drafting clinical notes.

You will review outputs of AI agents against clinical documentation to ensure that codes are supported, billable, and accurate. This role emphasizes critical thinking and defensible reasoning over high-volume processing; you will help shape guidelines, define clinical metrics, and participate

Qualifications

• Experienced medical coder with proficiency in ICD-10-CM, CPT, and HCPCS.
• Ability to review AI-generated codes against clinical documentation for support and billability.
• Demonstrated critical thinking to improve systems and define clinical metrics.

Responsibilities

• Review AI-generated ICD-10-CM, CPT, and HCPCS code sets against the clinical documentation, recording for each code whether the note supports it and where.
• Distinguish clinically defensible additions from unsupported inference. Identify both over-coding and under-coding.
• Evaluate E/M level selection, including correct setting and visit type, against documentation and CMS guidelines.
• Apply and improve the scoring guide consistently from the first case to the hundredth.
• Participate in dual review, independently scoring the same encounter and resolving disagreements with documentation.

Skills

Medical coding

ICD-10-CM

CPT

HCPCS

Code review

Quality review

CMS guidelines

Job description

Position Summary

Veradigm is seeking experienced Medical Coders to support our enterprise coding capability. Applicants will focus on ensuring accuracy in AI-generated medical codes, chart preparation, and drafting clinical notes. You will review the outputs of AI agents and compare to clinical documentation, to ensure that codes are supported, billable, and accurate. Coders will also shape the guidelines we use and help define clinical metrics we report to customers.

This is a role where you will use your experience and critical thinking skills to improve systems. It is not a standard coding position; you will not handle billing, provider education, or denials. The consistency and defensibility of your reasoning are prioritized over high-volume processing.

Key Responsibilities

• Coding Evaluation
  • Review AI-generated ICD-10-CM, CPT, and HCPCS code sets against the clinical documentation, recording for each code whether the note supports it and where.
  • Distinguish clinically defensible additions from unsupported inference. Identify both over-coding, which creates audit risk, and under-coding, which leaves care unreimbursed.
  • Evaluate E/M level selection, including correct setting and visit type, against documentation and applicable CMS guidelines.
  • Apply and improve the scoring guide consistently from the first case to the hundredth.
  • Participate in dual review, independently scoring the same encounter as another reviewer and working through disagreements to a documented resolution.
  • Flag cases where the scoring guide is ambiguous or silent rather than resolving them quietly, so the guide can be corrected before the ambiguity spreads through the dataset.
  • Work across specialties, beginning with internal medicine, applying specialty-specific conventions and noting where documentation patterns change how output should be judged.
• Beyond Coding
  • Review outputs that providers flag as incorrect in our products, determining whether the flag reflects a genuine agent error, a documentation gap, or a difference in provider preference.
  • Help validate automated scoring systems by reviewing a sample of cases an automated judge has already scored, so we can measure where it agrees with expert judgment.
  • Explain patterns in agent errors to machine learning engineers and product managers in terms they can act on, escalating systemic issues rather than scoring the same error repeatedly, and give feedback where the review tool slows accurate review.