Portfolio · 2026
Software engineer designing and shipping reliable systems end to end.
I work across API and data contracts, asynchronous service workflows, frontend integration, CI/CD pipelines, observability, and runtime performance tuning. I select tools based on constraints and tradeoffs, then optimize for correctness, latency, and reliability in production.
Choose tools for the problem. Design for reliability. Execute with discipline.
I start with system constraints first: latency budgets, failure domains, ownership boundaries, and data guarantees. Architecture decisions follow these constraints, not trends.
I translate ambiguous product asks into deterministic execution paths, measurable system behavior, and repeatable delivery so teams can move quickly without losing reliability.
I do my best work at the intersection of distributed infrastructure, applied AI workflows, and product engineering. I am intentionally stack-flexible and choose technologies based on constraints, tradeoffs, and the fastest path to reliable outcomes.
Scalable Python framework for production-grade AI agents
Provides a highly extensible agent runtime for single-agent and multi-agent orchestration, complex tool calling, and configurable pre-hooks, post-hooks, and middleware chains for each tool execution path.
Includes built-in memory primitives and DX-first APIs to ship adaptable agent workflows faster.
AI-powered human-like mouse and keyboard automation
A Python package for realistic, human-like input automation that predicts natural cursor trajectories and generates smooth mouse movements with adaptive noise and speed variation.
Includes smart keystroke simulation with human-style delays, punctuation-aware typing, and variable timing for organic interaction flows.
High-throughput TypeScript service runtime
Implements an optimized request lifecycle and routing path to reduce overhead in high-concurrency API workloads.
Performance wins only matter when maintainability and developer ergonomics remain strong.
Custom compact transformer model engineered for local experimentation
Designed and trained a 33M-parameter transformer locally with a self-trained BLE tokenizer, built to validate end-to-end model design decisions under tight compute constraints while preserving practical context depth.
Training profile: block size 1024, learning rate 3e-4, embedding size 348, 6 attention heads, and 12 transformer layers.
Dec 2022 — Jul 2024
Applied Engineering
Built distributed CRM and exam systems with RabbitMQ, Express, React Native, PostgreSQL, and Docker across regional deployments.
150% throughput gain · 40% faster CI/CD deployments
Aug 2024 — May 2028
University of Southern Mississippi
Focused on distributed systems, algorithms, and production software engineering.
4.0 GPA · Google Generation Scholar 2024
Public GitHub activity from May 2025 to May 2026, representing engineering execution cadence across repositories, pull requests, and shipping activity.
Open to software engineering roles, high-impact technical collaboration, and product/system design conversations.