Started as a CSDN technical column · Led by Dr. Xie

Medical Data Science
Meets AI Agents MEDICAL DATA SCIENCE × AI AGENTS

We work on medical data mining, clinical prediction models and enterprise AI agents; from statistical analysis to agent engineering, delivering reproducible, production-ready solutions. Our own tools: the macodgent R package and the RAI R AI IDE.

Browse Columns → Get in Touch
Who We Are

About Us

MA Codgent is led by Dr. Xie and grounded in medical data science methodology, spanning clinical prediction models, statistical analysis and AI agent engineering. We deliver for university research groups and enterprise clients, with medical rigor and reproducible methods as standard. Our own tools include the macodgent R package and the RAI R AI IDE.

We believe AI makes writing code cheap; knowing whether an analysis is correct is what creates value. Medical review and reproducibility are the baseline for everything we publish and deliver.

Our flagship column Exploring Medical Data Science with R is the core teaching text for partner research groups, videos and on-site training; companion columns on applied AI in medical research and AIGC engineering are in progress.
MA Codgent team poster
Team poster · Download high-res ↓
Our Software

Methodology, Shipped as Software

Alongside content and custom delivery, we build and maintain two tools in-house: the macodgent R package and the RAI R AI IDE. Both are publicly available.

R PACKAGE · V1.1.0

macodgent

62 exported functions · MIT · Open source

Reusable functions for the statistics, modelling and visualization workflows of routine medical data analysis. Methods and code are fully auditable, ready for research delivery and teaching.

remotes::install_github("XieYj2026/macodgent")
View source repository →
R AI IDE · MACOS / WINDOWS

RAI

A local workspace for R analysis

A local IDE built for medical research: persistent R sessions, editable code, an AI assistant and document preview, with support for your own R packages and plugins. Every step stays traceable.

Explore RAI →
Columns

Structured Columns, Built for Practice

From R for medical statistics to AI-assisted research, systematically organized, reproducible and continuously updated. Open a card to subscribe or purchase.

Content Matrix

Find Us Where You Already Read

One body of work, distributed across platforms. Follow us wherever is convenient.

Two Service Lines

Two Service Lines

Two types of delivery for university research groups and SMEs, from methodology to production engineering.

LINE ①

Clinical Prediction Models and Patents

  • Model development (nomograms / machine learning)
  • Publication-grade validation and manuscript support
  • Patent novelty mining and filing support
  • Reproducible end to end, hands-on support
LINE ②

Enterprise AI Agents

  • RAG knowledge bases / workflow automation
  • Multi-agent collaborative delivery
  • Lightweight builds for research groups and SMEs
  • Enterprise engineering standards, built to last
See the full custom program →
Custom AI Agents

AI Agents Built Around Your Business

Industry-agnostic: any setting with a knowledge base, repetitive workflows and human review. Medicine is our strength, not a requirement. Three hard requirements set us apart when serving government, finance and healthcare clients.

STARTER
Discovery Pack
Request a Proposal
~5 business days · scoping / use-case screening
  • 2 discovery sessions + workflow mapping
  • Ranked use cases with ROI estimates
  • Model cost forecast + implementation roadmap
  • A written Needs Assessment Report
Low commitment: decide if it is worth doing
VALIDATE
Validation
Request a Proposal
3–4 weeks · small-scale validation
  • 1 agent + knowledge base (≤300 docs / 2GB)
  • 1 automated workflow + web entry point
  • ≤50 test cases, 15-day warranty
  • Specification / acceptance report / user manual
Our recommended first engagement: visible value
PRODUCTION
Production
Request a Proposal
6–9 weeks · one department live
  • Production agent + up to 3 workflows
  • Knowledge base (≤500 docs / 5GB, traceable)
  • Monitoring dashboard + ≤100 test cases, 30-day warranty
  • Technical design / ops manual / training
Transitions to annual maintenance after launch

Pricing is scoped to project size and complexity · Request a custom proposal →

E-commerceEducationFinanceGovernmentManufacturingHealthcareLegalMedia
Add-ons

Optional Modules

Base pricing stays fixed; we add services rather than discount, protecting the brand and margin.

On-premise deployment
Optional
Client provides the local inference stack; data never leaves the site.
Operator training
Optional
Hands-on operations training built on our applied AI handbook.
Medical review and compliance
Optional
For healthcare and pharma clients; final sign-off by an MD.
Messaging integration
Optional
Feishu / WeCom / DingTalk integration.
Human approval gates
Optional
Manual confirmation for critical actions.
Single sign-on (SSO)
Optional
Connects to your corporate identity system.
Model cost optimization
Optional
Lower cost per inference call.
Domestic LLM connectivity
Included
No separate charge; part of every engagement.
Capability Matrix

Capabilities

Clinical, statistics and AI engineering, covering the full research lifecycle.

01
Clinical prediction models
Competing risks · nomograms · calibration · DCA
02
Patent novelty mining
disclosure → compliant translation
03
Enterprise AI Agents
RAG · workflows · multi-agent
04
Medical data governance
Statistical and clinical dual review
05
R for medical statistics
Cox · visualization · meta-analysis · ML
06
Publication-grade validation
External validation · TRIPOD
07
Reusable Skills
NHANES replication · science writing
08
Content distribution
Compounding authority across platforms
Let's Work Together

Have a Project in Mind?

Whether it is clinical prediction models, publication-grade methodology or enterprise AI agents, we would like to hear from you. Professors and research groups are also welcome to discuss teaching and training.

WeChat Jian_Kayla
Visit our columns