2024 - now
Associate Software Engineer, McKesson
Working on enterprise software with a focus on AI-assisted support tooling, data workflows, and production reliability.
Education
I’m a BS in Computer Science student at The University of Texas at Dallas, focused on AI/ML, distributed systems, databases, and production software engineering.
Current Work
2022 - 2024
Developer I, United Airlines
Worked on distributed payment and ticketing services, event-driven workflows, database migration, and production systems supporting airline operations.
Selected Projects
GleanJobs
A distributed job-search stack for software, AI, systems, and QA roles, refreshed daily from hundreds of company career boards.
OraFlow
A local evidence layer for database and Jira investigations: target binding, schema lookup, read-only SQL, audit trails, and packaged VS Code/MCP workflows.
compuse
A Windows computer-use MCP server that observes UI elements and drives real mouse and keyboard actions with guardrails.
MeetingAssist
A local meeting copilot for macOS and Chrome with live audio capture, OpenAI Realtime Whisper transcription, GPT-4o fallback, and AI chat over transcripts.
pAIralegal
An AI insurance-regulation assistant combining SERFF scraping, document processing, embeddings, Qdrant vector search, and LLaMA-based answers.
LM Go
A SwiftUI iOS chat client for LM Studio, OpenAI-compatible APIs, and on-device MLX inference with streaming, model controls, and local model discovery.
Flood Risk Prediction
A geospatial ML risk model using property assessment data, flood probability data, CatBoost, GeoPandas, and expected-loss analysis.
OrchardSDK
A TypeScript SDK for trusted agent-human workflows with threads, approval-gated actions, append-only audit events, idempotency, and subscriptions.
AlgoScholar
A Claude and Streamlit research assistant for ARXIV search, summarization, paper-network exploration, and research chat sessions.
Notes
- AI/ML systems should stay grounded in retrieval, logs, evals, and operator-visible evidence.
- Distributed systems need boring reliability: idempotency, audit trails, retries, and clear failure modes.
- I prefer fast, inspectable tools over unnecessary platform weight.
Contact
The easiest place to find my work is github.com/axm0. For direct notes, email [email protected].