I’m Naveen Benjamin. I work alongside business and technical teams to turn claims data, operational rules, and applied AI into useful systems.
I work across discovery, modeling, implementation support, and impact measurement. These examples describe the kinds of problems I lead and build for.
Designed an “automation factory” approach: detect high-volume opportunities, diagnose system behavior, align with subject matter experts, translate policy into rules, and measure results after implementation.
Built diagnostic frameworks, production datasets, predictive analysis, and capacity projections. The multi-team program contributed to a 60%+ reduction in processing cost over four years; modeled avoided operating cost is approximately $80M annually. These are program outcomes, not individual attribution.
Prototyped policy-grounded rule extraction and claim narratives, and designed a four-agent LLM architecture. AI explains patterns for expert review; deterministic rules govern final claims decisions. Authorization mismatch agents are in progress.
Public code and self-directed projects. Client systems and confidential claims data stay out of public repositories.
A local-first event photography assistant that groups near-duplicate frames, recommends selections using explainable visual signals, and prepares approved RAW copies with per-photo Lightroom XMP settings. Photographers review the choices; the source images stay untouched.
Explore the code on GitHub ↗An automated video pipeline for Wide Lens content, organized into Discover, Review, Produce, and Final Reel stages. The project brings those stages into a single workflow.
Private code; described here as a personal project.Other personal builds: Stratiic, an AI-assisted proposal workflow; ContextKit, a three-layer Markdown memory architecture; and OpenClaw Mission Control, a seven-agent orchestration system. These are self-directed projects, not deployed client products.
My work lives alongside claims operations, analytics teams, and leadership. I establish the data and deterministic rules first, then prototype AI diagnostics with human review. Claim payment decisions remain in governed business systems.
Embedded in Kaiser Permanente’s Health Plan Analytics & Insights team. Grew from individual contributor work into leading the analytics and AI layer of a multi-team claims automation program. Built pattern-mining and forensic analytics capabilities, translated payment logic into auditable rules, and framed LLM-assisted diagnostics. Authorization analysis agents remain in progress.
Worked across healthcare, retail and consumer goods, and financial services clients. Delivered data engineering, reporting, and analytical solutions.
MBA — Madras University (2010). B.Sc. Computer Science — Bharathiar University (2008).
Claims adjudication, payment integrity, contracts and fee schedules, prior authorization, Epic Clarity, EDI, and X12 837.
Python, SQL, PostgreSQL, Databricks on Azure, Delta Lake, data pipelines, Tableau, and Power BI.
Predictive modeling, capacity forecasts, LLM workflows, RAG patterns, agent design, and human review.
Cross-functional discovery, hands-on prototyping, team mentorship, stakeholder alignment, and executive communication.
I’m interested in principal, staff, and director roles across GenAI, ML and data science, and analytics leadership.