Integration and interoperability
Transforming structured healthcare data into dependable API workflows, with validation and mapping made explicit.
Healthcare integration · practical automation
Reliable software for work that needs to hold up.
Healthcare integration engineer building secure, testable automation and reliable data workflows with Python, SQL, APIs, and C#.
I build systems that turn complex interfaces and repetitive operational work into maintainable, reviewable software. My public projects emphasize explicit boundaries, deterministic behavior, careful failure handling, and evidence over hype.
Engineering focus
Useful systems, deliberately engineered
I focus on the places where correctness, clarity, and operational trust matter more than novelty.
Automation with guardrails
Replacing repetitive work while preserving confirmation, observability, recovery paths, and human control.
Evidence-driven quality
Using automated tests, static analysis, CI, documentation, and narrow boundaries to support claims about behavior.
Featured projects
Engineering choices you can inspect
Each project addresses a practical problem and documents both what works and where its boundaries remain.
01 · Security-minded application design
SQL Password Locker
A SQL Server-backed Python password vault that encrypts credential passwords before persistence and exposes the same behavior through a CLI and responsive Tkinter interface.
Engineering highlights
Layered services, transactional parameterized persistence, versioned schema setup, enforced encrypted transport, bounded background work, and isolated automated checks.
Honest boundary
This portfolio project has not received a professional security audit. It does not provide master-password recovery, key rotation, or protection for a fully compromised host.
Read the video transcript
The SQL Password Locker opens to an unlocked, empty encrypted credential vault. A synthetic account named “portfolio-demo” is entered with a demonstration password and saved.
The credential appears in the account list, demonstrating encrypted persistence. The password is copied to the clipboard, and a message confirms that the clipboard will clear automatically after 30 seconds.
The synthetic credential is deleted, returning the vault to an empty state. Finally, the vault is locked and the interface returns to the master-password unlock screen.
02 · Safety-first developer tooling
Project Creation Automation
A cross-platform Python CLI for confirmed local Git project creation and optional private-by-default GitHub repository creation.
Engineering highlights
Redacted plans, default-no confirmation, allowlisted commands, narrow adapters, credential-safe diagnostics, and conservative rollback keep every mutation reviewable.
Attribution and boundary
This GPL-3.0-or-later project is a substantial security-focused modernization of Tim Eichinger’s Windows implementation, which was inspired by Kalle Hallden’s original project-automation concept. The original concept is not claimed as independent work.
Read the video transcript
0:00–0:10 — Preview the plan. The plan command previews creation of “portfolio-demo-project” with --ide vscode. The destination is redacted, GitHub creation is not requested, and mutation_performed: no confirms that the dry run makes no changes.
0:10–0:16 — Confirm creation. The create command presents its plan and asks “Proceed with the planned creation? [y/N]”. Entering y approves local project creation.
0:16–0:21 — Check the result. The output reports confirmation accepted, directory and starter files created, Git initialized, starter files staged, the index verified, the initial commit created, and the IDE launched. The operation succeeds with a local project; GitHub creation is not performed. This result frame is held for readability.
0:21–0:28 — Start building. VS Code shows the generated project with README.md and .gitignore. The README contains the project heading and a placeholder description. The ending is held for readability.
03 · Deterministic data workflow
NFL Pool Automation
Python automation that retrieves public point-spread data and updates a structured Excel pool workbook while preserving stable matchup rows.
Engineering highlights
Retrying data ingestion, explicit team normalization, Pacific-time schedule handling, deterministic row identity, and saved fixtures for unusual weekly schedules.
Honest boundary
The scraper depends on third-party markup, workbook writes are not transactional, and live website, notification, and production-workbook integrations are outside the automated suite. The project is not affiliated with or endorsed by the NFL.
04 · Privacy-conscious desktop utility
Morning App Launcher
A Windows desktop utility for saving a short application set and opening the right tools at the start of the day.
Engineering highlights
Layered controller and presentation logic, safe legacy migration, injectable process launching, keyboard-accessible interaction, and logs that omit application paths.
Honest boundary
The application launches user-selected files with the current user’s permissions. It does not sandbox processes, scan files, sign executables, or replace Windows security controls.
Read the video transcript
The Morning App Launcher opens with an empty morning application list. The Add application action is used to choose an application and add it to the morning list.
The first application’s technical filename is replaced with a readable display name, making the saved entry easier to recognize. Another application is then added and given its own readable name.
With both applications ready, Open all applications is selected once. The launcher opens the complete morning application list with that single click.
05 · Deterministic legacy modernization
Hyphy Oregon Conference Generator
A cross-platform .NET CLI that divides fantasy-football league owners evenly between conferences through interactive or scripted workflows.
Engineering highlights
Core and CLI separation, balanced assignment, stable seeded output, Unicode-aware input validation, explicit exit codes, and portable release packaging.
Honest boundary
Seeded output is deterministic, not cryptographically random or externally certified for fairness. The project is independent and claims no league or organizational endorsement.
Read the video transcript
0:00–0:06 — Choose the setup. The default setup assigns ten owners evenly between the East and West conferences. Fictional names are used throughout the demonstration.
0:06–0:14 — Enter the owners. Alex, Blake, Casey, Devon, Emery, Finley, Gray, Harper, Indigo, and Jordan are entered in a consistent order. Name entry is accelerated.
0:14–0:19 — Set a reproducible seed. Seed 20200830 is supplied. The same setup, owner input order, and seed reproduce the draw.
0:19–0:30 — Review the conferences. East contains Indigo, Devon, Casey, Finley, and Blake. West contains Emery, Jordan, Alex, Gray, and Harper. The generator identifies its deterministic algorithm as SplitMix64-v1, and the recorded result is held for readability.
Project evolution
Modernization as an engineering discipline
The most useful before-and-after stories are about clearer boundaries and safer behavior—not simply newer tools.
SQLite Password Locker → SQL Password Locker
SQLite Password Locker established an authenticated local vault with scrypt, AES-256-GCM, SQLite, a CLI, and a Tkinter interface. The SQL-backed successor preserved the application-layer encryption model while introducing Argon2id key wrapping, transactional ODBC persistence, schema management, validated TLS transport, and a bounded background worker for the GUI.
The progression is not a claim that either project is professionally audited. It is evidence of increasingly explicit threat models, infrastructure boundaries, and operational limitations.
Preserve intent, replace fragile boundaries
Project Creation Automation and Hyphy Oregon both retain meaningful project history while replacing tightly coupled legacy behavior with testable layers. One focuses on safe filesystem, Git, GitHub, credential, and IDE boundaries; the other separates assignment rules from terminal interaction and defines a stable seeded generator.
The shared lesson is that modernization works best when behavior, provenance, limitations, and migration decisions remain inspectable.
Supporting work
Learning through teaching
Documentation for mixed-experience audiences
Git Cheat Sheet
I originally created the Git Cheat Sheet for FreeCodeCamp’s technical documentation project while participating in a Portland-area Python group that wanted to learn more about Git. I turned the assignment into a practical command reference for that group. Over time, it also became a useful resource for support coworkers who needed Git guidance without working deep in application development.
It remains a practical quick reference rather than an authoritative or complete source of Git documentation. Its lasting value is the exercise of explaining potentially destructive tools clearly to people with different levels of experience.
Review the Git Cheat Sheet on GitHubPractical Python learning reference
Python Virtual Environment Guide
I created this focused guide as a reusable explanation of Python virtual environments. It complements the Git Cheat Sheet by turning a common development setup task into documentation that can support people with different levels of Python experience.
It is supporting educational work rather than a featured software system, but it reflects the same emphasis on clear steps, practical context, and documentation that remains useful after the original learning exercise.
Review the Python Virtual Environment GuideEvidence-based skills
Capabilities reflected in public work
These groupings describe technologies and practices demonstrated by the projects above and the verified professional summary below.
Integration and data
- HL7 v2 and EMR workflows
- REST APIs, JSON, XML, and XSLT
- SQL Server, SQLite, ODBC, pandas, and Excel automation
- Validation, mapping, transformation, and time-zone handling
Languages and interfaces
- Python, SQL, and C#
- Command-line application design
- Tkinter desktop interfaces
- Semantic HTML and responsive CSS
Quality and delivery
- Automated tests and dependency isolation
- Static analysis, type checking, and formatting gates
- GitHub Actions and release packaging
- Technical documentation and explicit limitations
Defensive engineering
- Authenticated encryption and key derivation
- Credential-safe configuration boundaries
- Transactional persistence and atomic file replacement
- Confirmation, rollback, and privacy-conscious logging
Professional background
Healthcare integration since 2017
Since 2017, my work has involved developing and supporting HL7 integrations that transform data from electronic medical record systems into product-specific API workflows.
That work spans requirements analysis, data transformation, automated testing, deployments, production support, and customer-facing troubleshooting. The through-line is practical: reduce repetitive effort, strengthen validation, and leave systems easier to operate and maintain.
Development process and authorship
Transparent about how this site was made
This site and its project narratives were developed by David Mevorah with AI-assisted research, drafting, and implementation. David selected the content, verified the claims, and retains responsibility for what is published.
Contact
Continue the conversation
Explore the source, project histories, and current public work on GitHub, or connect professionally on LinkedIn.