Vishwajeet Bhabra
Basking Ridge
Summary
Senior AI/ML Engineer and AI Initiative Lead with 9+ years at Optum, specializing in enterprise AI systems, conversational multi-agent platforms, and HIPAA-compliant healthcare automation. Currently leading the end-to-end design and development of the DSS Genie AI Chatbot platform — a production-grade multi-agent AI system built on Azure OpenAI (GPT) that powers intelligent test data generation, EDI 837 claim creation, member management, and natural language data querying for the Cirrus healthcare platform. Deep expertise in Python, Java Spring Boot, LLM orchestration, prompt engineering, and zero-trust AI security architecture. Proven record of delivering high-impact AI solutions that dramatically reduce manual effort, improve data quality, and accelerate testing cycles in complex enterprise healthcare environments.
Overview
10
10
years of professional experience
1
1
Certification
Work History
Sr. Software Engineer & AI Initiative Lead
Optum
Basking Ridge
01.2022 - Current
- Lead the end-to-end design and development of the DSS Genie AI Chatbot — a multi-agent AI platform built on Azure OpenAI that enables test teams to create members, member groups, and HIPAA-compliant EDI 837P/837I claims through natural language conversations across all enviroments.
- Architected the AI Orchestrator layer — central LLM brain that classifies 11 distinct user intents and dynamically routes to the correct domain agent: Claim Agent, Member Agent, Member Group Agent, Data Mining Agent, Validation Agent, and Fallback Agent.
- Designed and built the Claim Agent: conversational HIPAA EDI 837P/837I builder with INN/OON provider auto-selection , CPT/ICD-10 code validation, bulk CSV-to-EDI batch processing, and one-click Cirrus upload — reducing claim creation time from ~30 minutes to under 1 minute.
- Implemented CPT/ICD-10 AI validation layer that checks HIPAA code compliance and CPT–diagnosis compatibility before EDI generation, significantly reducing Cirrus claim rejection rates during testing.
- Implemented INN/OON provider auto-search: AI derives POS-specific provider taxonomy and queries rso_01 in real time to select the correct in-network or out-of-network provider, eliminating all manual provider lookup for INN and OON test scenarios.
- Built the Chatbot Agent layer handling member creation, family/dependent enrollment, member group setup (book-of-business, reporting code, state), and natural language SQL generation across 6 entity types — enabling non-technical testers to create and query DSS test data without any SQL knowledge.
- Developed AI Rally Agent: LLM-powered exception analyzer that automatically determines severity, generates root cause descriptions, and creates/deduplicates Rally defects, ensuring 100% exception traceability without manual defect entry.
- Integrated RAG conversation memory (same-intent and cross-intent history injection) enabling context-aware multi-turn dialogues — users never need to repeat already-provided information across a session.
- Secured AI platform using mTLS (Spring Boot ↔ Python Flask microservice), OAuth2 token management for Azure OpenAI calls, and Service Principal-based Azure AI Search in read-only mode to enhance data protection.
- Designed and delivered bulk claim processing: paste raw CSV or free-text → LLM extracts all claims → validates → builds EDI → one-click upload; enabling 100's claims to be processed in a single chat interaction.
Software Engineer
Optum
Basking Ridge
01.2018 - 12.2021
- Developed new automation tools for Group, Member, and Claims domains; collaborated with stakeholders to adapt to evolving business requirements.
- Developed automation tools for claims creation for users with limited X12 (EDI) knowledge; implemented EDI Claim Load solution for faster claims processing and comprehensive reporting.
- Enhanced DSS by adding Claims Category and Claims Form features; streamlined authorization and referrals loads using JSON and Excel formats.
- Ensured HIPAA compliance with EDI 834 and 837 validations across all claim types.
- Developed Smart Database Genie, a comprehensive data search solution for Cirrus Platform, laying groundwork for AI-powered DSS Genie platform.
- Operated within a DevOps / Agile model with continuous delivery and daily deployments.
Technical Development Program
Optum
Basking Ridge
06.2016 - 12.2017
- Developed 83x File Generator tool (Python / MySQL / PyQT) for producing valid, ready-for-adjudication 837 and 834 claims from existing templates.
- Created Python script for FACETS Claim Solution to merge multiple 837 files, strip X12 headers/footers, differentiate institutional vs. professional claims, and validate HIPAA compliance before Facets submission.
- Developed VBA-based data reconciliation utility for 7 downstream applications (BRMS, PH-Accent, Optum RX, Dental, Vision, BIS, Evicore) to ensure data parity across systems.
Education
Bachelor of Science - Computer Engineering
New Jersey Institute of Technology (NJIT)
Newark, NJ12-2014
Advance Generative AI - Artificial Intelligence
Purdue University
West Lafayette, INSkills
- AI / LLM
- Azure OpenAI
- Claude
- Codex
- LangChain
- RAG
- Prompt Engineering
- Intent Classification
- NLP
- Azure Cognitive Services
- Azure AI Search
- Azure AI Services
- Azure Bot Service
- Multi-Agent Orchestration
- Python
- Java
- SQL
- Oracle
- Flask
- Spring Boot
- HTML
- CSS
- VBA Scripting
- Unix Shell Scripting
- NumPy
- Pandas
- Docker
- Kubernetes
- OpenShift
- Jenkins
- Git
- Amazon S3
- Apache Maven
- Tomcat
- MongoDB
- MTLS
- OAuth2
- HIPAA EDI 834
- HIPAA EDI 837
- Rally
- Tricentis TOSCA
- Sauce Labs
- Microsoft Office
Accomplishments
- Led the AI transformation of the DSS platform — evolved from a manual form-fill tool to a fully conversational, multi-agent AI system handling claims, members, and data queries end-to-end.
- Reduced HIPAA EDI claim creation time from ~30 minutes to under 1 minutes through AI-guided automation.
- Enabled bulk processing of 100s claims in a single interaction, compressing multi-hour testing cycles into minutes.
- Achieved 80% testing time reduction on enterprise automation initiative using Tricentis TOSCA.
- Designed a zero-trust AI security model (mTLS + OAuth2 + SP auth) for HIPAA-sensitive AI infrastructure in non-production environments.
Certification
- Advance Generative AI Certification from Perdue University
- AI Dojo
Timeline
Sr. Software Engineer & AI Initiative Lead
Optum
01.2022 - Current
Software Engineer
Optum
01.2018 - 12.2021
Technical Development Program
Optum
06.2016 - 12.2017
Bachelor of Science - Computer Engineering
New Jersey Institute of Technology (NJIT)
Advance Generative AI - Artificial Intelligence
Purdue University