About Me
Software engineer with a background in research, now building software systems and AI-forward products.
Contact Information
- Location: Austin, TX, US
- Email: [email protected]
- Phone: (512) 666-1737
- Website: desertthunder.dev
Education
Master of Science, Computer Science
Bachelor of Science, Microbiology
Bachelor of Science, Biochemistry
Skills & Expertise
Languages
Python, TypeScript, JavaScript, Ruby, Golang, SQL, Rust, Dart, ML Languages (OCaml, F#, Gleam)
Technologies
Django, React, React Native, Flutter, Svelte, Vue, Rails, Node.js, Docker, Kubernetes, ElasticSearch, Postgres, AWS, GCP
AI
LLMs, Model Context Protocol, Agent Client Protocol, Anthropic & OpenAI APIs, Agent Context & Skills
Professional Experience
Stormlight Labs
Founder and Lead Software Engineer
2024 – Present
- Founded and led an open-source-focused software studio, taking products from concept through production across multiple platforms.
- Turned research, user feedback, and complex product requirements into maintainable software with clear, practical user workflows
- Shipped and tuned multi-platform applications and open-source libraries in Svelte, Rust, Go, and Flutter, using concurrency design and profiling to improve reliability and runtime performance
Edlink
Senior Software Engineer
May 2023 – Jul 2024
- Led Flow as a production data pipeline connecting learning-management and student-information systems for school roster data.
- Extended Flow’s data model and API contracts to support new LMS and SIS integrations
- Designed an event-driven Pub/Sub, Cloud Functions, and BigQuery architecture that made school roster data durable, queryable, and easier to process across integrations
- Ran production roster-data workloads across Pub/Sub, BigQuery, Cloud Functions, Kubernetes, Redis, and Postgres, giving school integrations a reliable path from source systems to downstream consumers
- Delivered the Flow service in NestJS and TypeScript with Postgres, Redis, and Kubernetes on Google Cloud, plus a Vue dashboard for pipeline operations
CampusOptics
Software Engineer
Mar 2022 – Mar 2023
- Owned feature delivery for a higher-education environmental health and safety SaaS, from customer and management requirements through release.
- Delivered React Native features and the API endpoints behind them, expanding mobile workflows for campus safety teams
- Improved Rails and Sidekiq services on AWS ECS by tuning container tasks, background jobs, and Postgres workflows, keeping campus safety operations dependable
- Kept the Rails monolith, Sidekiq and Redis jobs, and Postgres services aligned on AWS ECS, providing a stable base for customer-facing releases
Texas Advanced Computing Center
Software Engineer
Oct 2019 – Mar 2022
- Architected and delivered the Frontera User Portal as the first implementation of Core Experience Portal v2.
- Delivered the portal across React, Django, and Postgres, turning the Core Experience Portal v2 design into a working user-facing system
- Improved DesignSafe-CI’s discoverability by developing an SEO strategy and updating its publication pipeline
- Introduced Pytest coverage for Core v2 and strengthened internal testing practices, giving the team a repeatable way to catch regressions
Projects
- Built a Rust terminal AI pair programmer with a custom TUI, streaming reasoning and tool activity, persistent session history, and project context loading from AGENTS.md and SKILL.md files
- Implemented provider adapters for Umans.ai and OpenCode Go with OpenAI-compatible and Anthropic-compatible tool schemas, model configuration, usage tracking, and provider-native or Exa-backed web search modes
- Designed bounded local tool execution, MCP tool discovery and dispatch, and Agent Client Protocol support so external agents, editors, and IDEs can drive the harness with auditable permission behavior
- Built a Rust compiler pipeline that parses Gleam with tree-sitter, owns its AST, resolves names, type checks a supported subset, lowers to core IR, and emits deterministic WebAssembly
- Implemented support for scalar values, managed runtime values, pattern matching, closures, selected standard library modules, host imports, and whole-project linked output
- Validated compiler behavior with source-span diagnostics, optional WAT rendering, and Wasmtime tests for scalar and managed-value exports
- Architected Rust-based LSP implementation with Hindley-Milner type inference engine
- Implemented comprehensive code intelligence including real-time diagnostics and type narrowing via control flow analysis
- Built refactoring suite with workspace-wide renaming validation and automated quick fixes
- Designed incremental analysis architecture with multi-layer caching supporting VS Code, Zed, Neovim, and Helix
- Built Flutter/Dart mobile client for iOS and Android with Material 3 design
- Implemented AT Protocol features including timelines, feeds, post composition, moderation tools, and account switching
- Designed Drift-backed local cache with offline-aware screens, drafts, saved posts, and semantic search
- Deployed a local-first SvelteKit application using PGlite and Drizzle, achieving sub-100 ms offline UX with deterministic data consistency
- Designed and implemented a premium sync engine with queued operations, SHA-256 diffing, and a network-aware worker
- Unified observability across web and worker contexts via session-scoped & GDPR compliant telemetry
- Built Zig Markdown engine with AST-first and pull-event APIs for renderers, formatters, and low-allocation integrations
- Implemented multi-dialect parsing with HTML rendering and URL sanitization controls
- Full CommonMark and GFM spec compliance with integrated conformance corpora
- Built Flutter/Dart browser with WebView-based text selection, highlighting, underlining, and note creation
- Implemented persistent annotations using layered W3C-inspired selectors for graceful anchoring across DOM changes
- Designed local-first state architecture with Riverpod, Drift/SQLite, go_router, and injected JavaScript bridge
- Built RESTful API with Go and PostgreSQL serving historical data from Lahman Database and Retrosheet
- Implemented advanced analytics including win probability calculations from 2,143 unique game states
- Designed trigger-maintained search vectors with GIN/trigram indexes enabling sub-millisecond queries
- Architected ETL pipeline with materialized views and comprehensive Swagger/OpenAPI documentation
- Architected terminal-native productivity system combining task management, note-taking, and media tracking
- Implemented repository pattern with SQLite persistence and Cobra CLI framework
- Built article parser with Colly and goquery for web scraping
- Built an Elixir/Phoenix PDS with account, repository, blob, access, security, migration, sequencer, and firehose operator tooling
- Implemented local development storage backed by SQLite databases and configurable blob limits, hostname, public URL, DID method, and crawler settings
- Maintained admin status and compatibility tooling for AT Protocol endpoint coverage, plus built-in documentation, changelog, and development email previews