Los Angeles, CA

Open to remote and hybrid roles

Kenny Santanu

Healthcare IT Project Manager | EHR and Revenue Cycle

I turn broken clinical and revenue cycle workflows into systems that work. Eight years inside a working clinic, building, shipping, and training people on the result.

Tell me your problem. I will find the technology and build the solution.

LinkedIn GitHub

$60K+

recovered in denied claims, year one of the AI denial tracker

80%

less time spent on manual billing after automation

30 days

EHR migration, zero downtime, 100% staff adoption

2,700+

PMI Los Angeles members served as Director of the Website

About

Operations first, technology second.

I am a PMP certified project manager with eight years inside a single private nephrology clinic in Los Angeles. That focused environment let me run revenue cycle operations, lead an emergency EHR migration, and design an AI powered denial tracker that recovered real money for the practice. I learned to manage projects by managing them, not just by reading about them.

My background blends healthcare operations with a working knowledge of software, data, and AI. I hold a Master of Science in Information Technology Management and a Bachelor of Science in Computer Science. I run AI agents in production for both clinical and personal work, and I write about what actually breaks when AI hits real workflows.

Today I am stepping into larger healthcare IT and EHR implementation roles, with a clear path toward Functional IT Manager. I am particularly drawn to organizations that sponsor Epic or Oracle Health certifications and to teams running clinical AI programs.

Credentials

PMP

Project Management Professional

Project Management Institute

CAPM

Certified Associate in Project Management

Project Management Institute

M.S.

Information Technology Management

Western Governors University

B.S.

Computer Science, Minor in Data Science

Cal Poly Pomona

Experience

Healthcare Revenue Cycle and Project Manager

Jonardi Handoko, M.D. • Upland, CA

Aug 2017 to Present

Designed and deployed an AI powered denial tracking system that recovered over $60,000 in underpaid and denied claims within the first year.
Led an emergency EHR migration in under 30 days with zero downtime and 100% staff adoption after a vendor data loss event.
Implemented billing automation that cut processing time by up to 80% and freed staff to focus on patient facing work.
Coordinate technology infrastructure refresh projects with zero service interruptions.
Manage end to end healthcare revenue cycle operations for the practice.

Director of the Website

PMI Los Angeles Chapter • Los Angeles, CA

Sep 2025 to Present

Maintain pmi-la.org with 99.9% uptime, serving more than 2,700 chapter members.
Run monthly training for new volunteers through documentation and hands on sessions.
Coordinate with vendor Dark Rhino on enhancements that improve member resources and engagement.
Partner with cross functional chapter teams on website projects from intake through release.

Projects

AI Denial Tracking System

Recovered $60,000+ in year one. Built and deployed inside an active clinic.

A progressive web app that streamlines investigation and resolution of insurance claim denials. The system gives next step recommendations, drafts appeal letters, checks payment status, and auto closes resolved denials. Deployed on premise so protected patient data never leaves the clinic.

My role. Designer, builder, and operator. Worked with billing staff to define workflows, then implemented the full stack and trained the team.

SvelteSkeleton UISupabaseAG GridTypeScript

Case Study

EHR Emergency Migration Case Study

30 day cutover, zero downtime, 100% staff adoption.

After repeated data loss events with the prior EHR vendor, I led an emergency migration from Aura by Sigmund Software to OfficeAlly EHR. Scoped vendor evaluation, data migration, parallel testing, training plan, and go live in under 30 days while keeping the clinic running.

My role. Project lead. Vendor selection, data mapping, training plan, change management, and cutover.

EHR ImplementationVendor ManagementChange ManagementTrainingData Migration

Case Study

Production AI Agent Workflow

Daily use of self hosted and frontier AI agents for clinic operations and personal productivity.

Architect and operate AI agent stacks using OpenClaw, local model runtimes such as LM Studio and Ollama, and frontier APIs. Hands on experience with prompt design, tool use, evaluation, governance, and the kinds of failure modes most healthcare teams have not yet seen in the wild.

My role. Architect and operator. Bridge between healthcare operations and applied AI.

OpenClawLM StudioOllamaPythonAI Governance

AI Discord Bot

Personal lab for testing emerging AI techniques in a sandboxed setting.

An ongoing experimental platform built with discord.py for integrating new open source and self hosted AI capabilities. Includes an LLM chatbot, image generation, text to speech, and other modules I rotate in as the field moves.

My role. Sole developer. Use this project to stay current on the model landscape and tool ecosystem.

PythonPyTorchOllamaDiscordDocker

SAM2 Image Segmentation and Replacement

Hands on integration of a frontier vision model into a working web app.

A Python web app that lets a user draw a rough mask, then uses Meta Segment Anything Model 2 to refine the mask for precise segmentation and seamless region replacement. Demonstrates comfort wiring deep learning models into real interfaces.

My role. Sole developer.

PythonGradioOpenCVMeta AIHugging Face

Case Study

YOLOv8 Traffic Sign Detection from Scratch

3,078 labeled images. Real dashcam footage. Trained without pre-trained weights.

A computer vision model built to detect and classify 17 categories of California traffic signs and signals, including stop signs, speed limits, and traffic light states. Collected hours of dashcam footage across seven SoCal cities, annotated every image manually in Roboflow, and trained YOLOv8 from scratch across three dataset iterations. The core finding: scaling data volume alone does not close the performance gap. Labeling consistency and architecture selection matter more. A final experiment with COCO pre-trained weights reached approximately 70% mAP.

My role. Co-researcher and primary dataset contributor. Personally annotated 1,344 images across all three dataset versions. Co-designed the training pipeline, ran local experiments on an RTX 3090, and co-authored the published research paper.

PythonPyTorchYOLOv8RoboflowGoogle ColabJupyter NotebookComputer Vision

Skills

Project Management

PMP and CAPM certified
Agile, Scrum, Waterfall, Hybrid
Full project lifecycle, initiation through closing
Cross functional coordination
Vendor evaluation and management
Stakeholder communication and business analysis

Healthcare

EHR implementation and migration
Revenue cycle management
Denial prevention and appeals
Clinical workflow design
HIPAA aware system design

Technology and AI

AI agents, prompt design, tool use, evaluation
Python, SQL, TypeScript
Local model inference with LM Studio and Ollama
Process automation
System integration and data migration
Jira, Trello, MS Office Suite, SharePoint

Have a problem worth solving?

I am open to mid level project manager and implementation roles in healthcare IT, revenue cycle, EHR, and clinical AI. Remote, hybrid in the Los Angeles area, or contract.

Message me on LinkedIn See work on GitHub