ABAnita B. Contact me

Health data engineerFounder

Anita B.

Building the data team that small clinics never had.

I work full time in health data at a large hospital system, where a quality number passes through engineers, analysts and reviewers before anyone acts on it. A small clinic owes many of the same numbers with far fewer people, so the work often lands on a nurse or an office manager after hours. I started MyCareFriend to close that gap.

zegore.com
Zego RE home page
bleanfoundation.org
Blean Foundation home page
mycarefriend.io
MyCareFriend home page: a full data and operations team, sized for your clinic
MyCareFriend on a phone
6Working papers on SSRN
128Paper downloads
71State organizations asked
4Data Notes for clinics
2Assistants ready for pilots

Paper counts from my SSRN dashboard, October 22, 2026. Outreach sent October 19 to 21, 2026.

Building

MyCareFriend is what I'm building. I also founded Blean Foundation and Zego RE.

The data team a small clinic could never hire

MyCareFriendFounded 2026 · mycarefriend.io

Hospitals have whole departments for data, finance, systems and quality. Most small clinics have never had that option, even though they carry much of the same reporting work.

When I went through HRSA's public data on federally funded health centers, the gap was plain. Of the 1,098 small and mid-size centers I ranked, 802 report no separate system for analysis and reporting, and 474 are rural.

So I'm building MyCareFriend as a small team of digital assistants, each made for one job, working alongside the people a clinic already has. Every number it produces shows its work: how many patients were counted, the months it covers, and how it was worked out.

Sample data for a made-up clinic: a dashboard showing diabetes control at 72 percent, 20 records to tidy and one late patient-list update
Sample data. The clinic is made up; this is the view MyCareFriend is building toward.

The first two assistants, an operations lead and a data analyst, already work end to end on realistic sample data. They read health records in the FHIR standard, run seven record checks, work out two quality measures for a clinic's own tracking, and flag an update that arrives late. The researcher, finance and systems assistants come next, then a patient companion. We are getting ready for our first pilot practices, and no patient information comes in until a signed business associate agreement, covered hosting and multi-factor sign-in are in place.

  • Ready for pilotsOperations lead and data analyst
  • In developmentResearcher, finance lead, systems and security lead
  • Coming laterPatient companion
Visit mycarefriend.io
bleanfoundation.org
Blean Foundation home page: Dignity. Education. Hope.
Blean FoundationFounded 2026

Blean is built on three words: dignity, education and hope. It is a Georgia nonprofit, tax exempt under 501(c)(3), launching programs in healthcare career pathways, free STEM courses for girls, and dignity services for people experiencing homelessness in Atlanta.

It is also where I plan to train the clinic staff who will review MyCareFriend's work.

Visit bleanfoundation.org
zegore.com
Zego RE home page: what if you could see inside a deal before you wired the money?
Zego REFounded 2025

Zego RE starts from a simple question: what if you could see inside a real estate deal before you wired the money? On most syndication platforms, the people selling the investment are the same people telling you it is safe.

Zego is built to score each deal separately from the sponsor selling it.

Visit zegore.com
Listening

October 2026: asking the people closest to small clinics before building more.

Before building more, I went and asked

In October I wanted to hear from the people who know small clinics best, before building anything else.

October 19 to 21, 2026
71

state organizations that work with small and rural clinics

41 state offices of rural health30 primary care associations

I sent each of them the two-page briefs of my four Data Notes, offered nothing for sale, and asked three questions. Replies are still coming in.

  1. Do these problems match what you see in your state's clinics?
  2. Which note is most useful, and what is missing?
  3. How much interest would clinics have in this kind of support?
What I'm working on next

Making it easy for a small clinic to say yes

  1. Stay with small and rural clinics. They have the least help today, which is the gap MyCareFriend was designed for.
  2. Answer the governance questions up front, in one readiness document: what MyCareFriend connects to, what it changes, how its checks were tested and how data is protected.
  3. Keep the clinic's review role small. Minutes a week, not a new job, and the first pilot will measure it.
  4. Train that role through Blean Foundation, with a short course for the staff member who takes it on.
  5. Find the money. I'm researching which public programs could help pay for this work, and on what terms, starting with the Small Rural Hospital Improvement Program and the Flex Program, both for small rural hospitals, and state plans under the Rural Health Transformation Program.
Writing

Open working papers, and short notes for clinics.

I write down what I learn

Every system I build leaves me with a question I want answered properly. So I write it up as a working paper and post it where anyone can read it. Three are design science case studies: build the system, test it honestly, and report what held up and what did not.

6papers
388abstract views
128downloads

Source: my SSRN My Papers dashboard, October 22, 2026

PaperDownloads

Bars show downloads per paper, on one scale. All papers on SSRN ↗

Data Notes for Small Clinics

For clinics, I write shorter.

Each note takes one common problem in clinic data and sets out steps a small team can follow, using synthetic records, simulated inputs or published evidence only. Each has a two-page brief. The briefs are free to use and share. Email me for the full notes.

  1. Three date checks to run before you count a quality measure
  2. Bringing scattered clinic records into one governed copy before you count
  3. Six rules for using language models in clinic data work
  4. Connected records and provider time: what the evidence shows
Tinkering

Systems I built for my own work and keep improving.

Two systems I built for myself

In use since April 2026

Grant-writing agents

A grant application can take weeks, and small organizations rarely have weeks to spare. So I built software that does the slow part: it searches federal and foundation funding, screens each opportunity for fit and drafts the application.

Six automated adversarial reviews pick each draft apart, one before it is written and five after, and then it comes to me. Nothing goes out until I have reviewed and approved it.

Find→Screen→Research→Review→Draft→5 reviews→My approval

I tested the method in a working paper with blind expert evaluation on 45 matched pairs, and it is how I research funding for MyCareFriend and Blean.

My longest-running project

A trading research system

This is where I test ideas about model evaluation and risk before I trust them anywhere else. Rule-based signal readers and a market-structure check vote on every trade. On gold and currency pairs, eleven machine learning models vote too.

When they disagree, a language model can veto that market and direction, based on its own read of the market. Stops and profit targets follow Fibonacci levels from recent price swings. The system itself trades only demo accounts.

11machine learning models
336features per model decision
4 or 5rule-based signal readers
Education

Degrees, and credentials I'm working toward.

I came to data through biology

Each step added a layer: biology first, then clinical prediction, and next, data management and AI engineering.

  1. 2030, anticipated

    D.Eng. in Artificial Intelligence and Machine Learning

    Planned research on agentic AI for clinic data quality.

  2. 2027, anticipated

    CDMP, Master level

    DAMA International's Certified Data Management Professional, covering data governance, data quality, architecture and modeling.

  3. 2024

    MS in Data Analytics · McDaniel College

    Machine learning for clinical prediction. My practicum built preeclampsia-risk classifiers on a hypothetical dataset. GPA 3.875.

  4. 2015

    BS · Southeastern Louisiana University

    Biology and pre-medicine coursework, including genetics and molecular biology.

Get in touch

If you work with small clinics, I'd like to hear from you.

Updated October 2026