Software engineer · Open to opportunities

I build software.
From design to delivery.

I design, implement, test, and ship software—from a published Chrome extension and a connectome-driven 3D game to data pipelines, Python applications, and responsive web interfaces. My work combines hands-on development with repository-scale review.

Explore my work ↘ Get in touch

BuildArchitecture & implementation

VerifyTests & code review

DeliverPackaging & documentation

Scroll

Experience

Software development.
Engineering ownership.

CURRENT ROLE

Software Engineer — AI Evaluation (Contract)

DataAnnotation Tech

Remote · 1+ year

Perform repository-level engineering reviews of AI-generated software changes. Analyze requirements, inspect implementations, write and assess tests, and diagnose defects to evaluate correctness, code quality, and functional behavior.

Code review: assess implementations in Python, Rust, JavaScript, and C++ against task requirements and surrounding code.
Testing & validation: write and assess tests, reproduce defects, and investigate edge cases and incomplete fixes.
Root-cause analysis: trace execution paths and interactions between components to explain failures at the source.
Functional verification: exercise applications and games to identify state-management defects, rendering issues, and usability problems.
Technical communication: document reproducible findings, compare implementation tradeoffs, and support conclusions with test results and code-level evidence.

INDEPENDENT PROJECTS

Software Engineering

Browser products · Python applications · web development

Architecture · Implementation · Delivery

Build and deliver browser products, Python applications, and web interfaces, owning the work from requirements and architecture through implementation, testing, packaging, and documentation. Public projects include a published website-diagnostics extension, a document retrieval application, an iPhone backup extraction utility, and this portfolio.

Product engineering: built and published SiteMed, a local-first Chrome extension with a diagnostic engine, saved comparisons, whole-site audits, availability monitoring, printable reports, and an automated release workflow.
Real-time systems: built Escape Circuit, a complete 3D game that connects live sensory inputs, connectome-derived controllers, fly biomechanics, collision, telemetry, audio, and replay.
Application delivery: developed and packaged RAG Pipeline’s retrieval engine with support for four document formats, ranked results with source citations, and command-line and browser interfaces.
Architecture & implementation: built modular ingestion, chunking, indexing, ranking, and persistence components, with shared application logic behind both interfaces.
Reliability: implemented validation for uploaded documents and saved indexes, atomic file replacement, and clear error handling to protect existing data when input is invalid.
Testing & CI: added unit and integration tests for search, storage, uploads, and CLI behavior; configured a Windows/Linux test matrix for Python 3.11 and 3.14 on pushes and pull requests.
Packaging & documentation: delivered an installable Python package, bundled browser assets, setup instructions, sample documents, and a technical walkthrough of the retrieval algorithms.
Workflow automation: developed a SQLite-backed backup extractor that automates file lookup and export across hashed and older flat iPhone backup layouts.

Selected work

Selected engineering work.

Browser products, applications, AI projects, and engineering contributions, with a focus on what I implemented and how the software works.

Gameplay capture Live circuit telemetry · first-person interaction · 3D pursuit

1.1 GBversion-locked connectome source data

39,549observed contacts in the escape artifact

≈3Mtriangles in the furnished environment

FLAGSHIP PROJECT · GAME & SIMULATION ENGINEERING

Escape Circuit / Fly Brain

Built a fully 3D first-person game in which compact controllers derived from measured fruit-fly connectivity influence escape actions and spontaneous turns. The project brings together real-time gameplay, neural-data processing, simulation architecture, testing, scientific documentation, and release engineering.

Complete gameplay system: implemented first-person movement, mouse tracking, charge-and-release swatter mechanics, timed rounds, scoring, pause/restart states, audio, telemetry, and a twelve-second split-screen replay.
Connectome data pipeline: built a reproducible Python workflow that verifies checksums, scans MaleCNS v1.0 source data, and generates compact, reviewable escape and flight-turn artifacts.
Real-time controller architecture: designed in-process neural providers plus an optional non-blocking Python reference service, keeping gameplay responsive even when external computation is unavailable or late.
Simulation engineering: translated modeled visual looming into distinct takeoff, directional escape, airborne-saccade, steering, power, and landing signals that drive the fly’s 3D state machine.
World & collision systems: adapted a detailed apartment into a navigable game space with shared furniture footprints, room-to-room waypoints, structural collision, movement substeps, and swept checks for fast-moving objects.
Quality & delivery: added deterministic Godot and Python tests, linting, type checks, data provenance, third-party attribution, and automated Windows and Web release packaging.

Engineering result: a self-contained game that runs without a server, account, or internet connection, while retaining an independently implemented Python path for verification and inspection.

Measured connectivity informs selected controllers; stimulus encoding, neural dynamics, action decoding, and biomechanics are explicitly documented engineering models—not a whole-brain simulation.

SiteMed Chrome extension showing a completed website scan and plain-English findings — Product interface Page evidence translated into clear next steps

ScanObserve the page
CompareTrack changes
ReportShare next steps

02 · PUBLISHED PRODUCT · CHROME EXTENSION ENGINEERING

SiteMed

Built and published a Chrome extension that diagnoses website problems and turns technical browser evidence into plain-English explanations and practical next steps for site owners, teams, and clients.

Diagnostic engine: implemented bounded checks for broken resources, accessibility gaps, form and download risks, trust signals, search setup, layout defects, and loading performance.
Extension architecture: separated React interfaces, a Manifest V3 service worker, an isolated content script, browser-independent diagnostic rules, and Chrome API adapters behind typed, runtime-validated messages.
Website care workspace: delivered saved scan comparisons, baselines, same-site audits, availability monitoring through Chrome alarms, client branding, printable reports, and JSON export.
Privacy and permissions: kept analysis and workspace data on the device, added no analytics or SiteMed server, and requested expanded site or notification access only when the user starts the related feature.
Quality and release: built unit coverage, a deterministic diagnostic fixture, a real-Chrome smoke flow, formatting, linting, type checks, production builds, and Chrome Web Store packaging.

Delivered: a published browser product with an official privacy and support site, a maintainable release process, and clear boundaries around what its evidence can establish.

Visit SiteMed

03 LOCAL DOCUMENT SEARCH

From files
to source evidence.

01
Load & chunk
PDF, DOCX, Markdown, and text
02
Index & rank
TF-IDF + BM25 scoring
03
Return the evidence
Original excerpts with source citations

Local processing · Inspectable ranking

Engineering decision

Used deterministic lexical ranking to keep search local, explainable, and testable. The CLI and browser share one retrieval pipeline, so fixes and improvements apply to both interfaces.

PYTHON APPLICATION · INFORMATION RETRIEVAL

RAG Pipeline

Built Grounded Search, the retrieval engine for my RAG Pipeline, to make local document collections searchable. Implemented the full path from PDF, DOCX, Markdown, and text ingestion to ranked passages with source citations.

Search implementation: built overlapping chunks, an inverted index, and configurable TF-IDF / BM25 scoring with reproducible ranking.
Feature delivery: implemented document uploads, index building and inspection, and search through a CLI and local HTTP interface.
Data integrity: added schema validation and atomic index replacement; validated uploads before replacing files and rejected unsafe filenames.
Automated verification: wrote tests for upload-to-search flows, overwrite handling, corrupt indexes, and CLI errors, with GitHub Actions configured for Windows/Linux and Python 3.11/3.14.
Developer handoff: packaged the application with its web assets and documented installation, usage, ranking algorithms, and example workflows.

Result: a local application that lets users upload documents, build a searchable index, and inspect original evidence through either interface.

View RAG Pipeline Technical walkthrough Test suite

04 PYTHON TOOLING · DATA EXTRACTION

iPhone Backup Extractor

Developed a Python utility that reads an iPhone backup’s SQLite manifest, resolves its internal file mapping, and exports Voice Memo recordings and the Notes database into organized folders.

Data access: queried the SQLite manifest to map original paths to the opaque identifiers used to store backup files.
Compatibility: resolved files in hashed subfolders and added a fallback for older flat backup layouts.
Export workflow: organized recordings and Notes database files into separate folders while preserving file metadata.
Error handling: checked for a missing manifest, skipped unavailable files, and reported extracted files and the final export count.

Result: automated manifest lookup and file export, turning opaque backup storage into organized audio and database files. Delivered with configuration and usage instructions using Python’s standard library.

View Backup Extractor

05 FRONTEND ENGINEERING

This Portfolio

Designed and developed this responsive website with HTML, CSS, and JavaScript, implementing the layout system, navigation, animation, and project presentation as a lightweight static application.

Frontend architecture: separated semantic HTML, a reusable CSS visual system, and JavaScript interactions into a self-contained static site.
Responsive behavior: implemented CSS Grid layouts, mobile navigation, and section-aware navigation highlighting.
Animation performance: bounded the canvas particle count and display scaling, and used IntersectionObserver for scroll reveals.
Accessibility: supported reduced motion, visible keyboard focus, skip navigation, labeled controls, and Escape-key menu dismissal.

Delivered: a responsive portfolio with project case studies and contact links, served as static HTML, CSS, and JavaScript without a build step or UI framework.

Discuss the build

06 PROFESSIONAL WORK · CODE REVIEW

Agentic Code Evaluation

Review AI-generated repository changes from the task requirements through to observed behavior. I investigate the surrounding code, reproduce failures, and assess whether the implementation and its tests support the claimed fix.

Trace failures across components to identify their root cause.
Compare proposed implementations and explain differences in correctness and quality.
Support evaluations with concrete test results and code-level evidence.

Professional work · Details generalized

07 PROFESSIONAL WORK · PRODUCT QUALITY

Interactive App Analysis

Evaluate applications and games as interactive systems, checking how interface behavior, state, and rendering line up with the requested experience.

Exercise user interactions to find broken state and inconsistent behavior.
Identify rendering defects and unclear interactions that affect usability.
Describe gaps between expected and observed behavior in structured defect reports.

Professional work · Details generalized

Capabilities

From architecture
to working software.

Languages

Python
GDScript
JavaScript / TypeScript
Rust
C++
SQL / SQLite
HTML & CSS

Engineering

Application architecture
Real-time systems & simulation
Automated testing & CI
Debugging & code review
Data pipelines & retrieval
Packaging & technical documentation

Tools

Godot 4
Chrome Extension APIs & WXT
Git & GitHub Actions
Docker
Playwright & Vitest
pytest & Python unittest
Python packaging & CLI tooling

Beyond the résumé

I take ownership
of the implementation.

I’m a software engineer focused on building useful applications and understanding the systems behind them. My project work spans browser products, real-time simulation, scientific data processing, document retrieval, automation, and frontend development.

I work through the full development process: breaking down requirements, implementing features, debugging failures, writing tests, and delivering code with usable interfaces and documentation. At DataAnnotation Tech, I apply that engineering judgment to reviewing changes in unfamiliar repositories and explaining their strengths and defects.

Outside of code, I’m interested in AI systems, game design, open source, language learning, and the way technology changes how people interact with the world.

AI systemsGame designOpen sourceLanguages

Have a role, project, or difficult problem?

Let’s build
dependable software.

I’m open to software engineering roles where I can contribute to feature development, testing, debugging, and the delivery of useful software.

Email me ↗ GitHub LinkedIn