Sandmark Technologies - Portfolio

Exploring neural networks, local models, and applications in agents, puzzles, and multi-agent systems. Includes integration with tools like reinforcement learning for complex behaviors.

Investigating swarm dynamics, area coverage strategies, and fault-tolerant simulations in sensor networks. SensorAnalysis on GitHub provides a platform for modular algorithm development using C# and reflection.

During my thesis and dissertation work, I gleaned many of my findings from the results I collected using this purpose-built simulation platform. Written in C#, the program heavily uses reflection and other abstraction techniques to allow dynamic, modular development of algorithms that govern the behavior and interaction of mobile sensors with their environment and each other.

Applying agile and lean methodologies to enhance team performance, avoid costly mistakes, and achieve predictable project outcomes with less stress.

I wrote an (as yet unpublished) book called Practically Agile that details my approach to implementing an Agile software development life cycle in a software team. After several successful commercial software development cycles using this approach, I wanted to share my experience and unlock key sticking points for developers and managers so they, too, could utilize the benefits without succumbing to the pitfalls.

An application developed with the Windows Universal Platform, published to the Windows store, designed to assist food pantry volunteers manage paperwork regarding community members served weekly with food and household items. GitHub Repo

I enjoyed playing Fallout 4. In certain situations, it presents users with a word guessing puzzle to unlock terminals. I wrote an AI program to assess the choices, recommend the best guess, evaluate the response of each guess, and iterate until the solution is found. This was a fun program written in C# that I then ported to iOS, Android, and WP8 using the Xamarin framework. GitHub Repo

During my thesis and dissertation work, I gleaned many of my findings from the results I collected using a purpose-built simulation platform. Written in C#, the program heavily uses reflection and other abstraction techniques to allow dynamic, modular development of algorithms that govern the behavior and interaction of mobile sensors with their environment and each other. GitHub Repo

A comprehensive .NET survey platform integrated with Couchbase for data storage, Docker for deployment, and JWT for authentication. Features UI components for survey modules and analytics dashboards. GitHub Repo

A MAUI-based Android app featuring a beach-themed user interface, integrations with health endpoints, and optimized image resources to boost user engagement and experience. GitHub Repo (assumed; confirm URL if different)

A Java-based service solving knapsack problems, comparing brute force and dynamic programming approaches for optimization and efficiency analysis. GitHub Repo (assumed; confirm URL if different)

Applying game theory to model AI development races, using replicator dynamics and interventions to achieve stable, sustainable equilibria. Includes LaTeX drafts on Overleaf and Jupyter simulations for multi-agent systems and market dynamics.

Quantifying trust envelopes in heterogeneous agent systems, analyzing the performance impacts of rogue agents in sparse networks. Draws analogies from FranklinCovey trust models, with simulations demonstrating resilience and degradation effects.

Investigating fault tolerance mechanisms and failure modes in sensor networks with limited mobility, focusing on maintaining coverage and connectivity through adaptive algorithms and simulations.

Practical applications of agile methodologies, emphasizing team management, iterative improvements, and balancing process with productivity in software development lifecycle.

Formulations and algorithmic solutions for detecting anomalous web traffic, including machine learning models and pattern recognition techniques for security and performance monitoring.

Managing DateTime calculations in Windows environments, addressing time zones, daylight savings, and edge cases to ensure accurate and reliable software behavior.

Setting up home labs with Home Assistant, local AI agents, camera integrations, Docker containers, and networking tools like Hyper-V Ubuntu VMs for experimentation and automation.

Tackling challenging CS problems like knapsack optimizations, recursive algorithms, and puzzles, with implementations in languages like Java and C# for efficiency comparisons.

A concept exploring two agent types: one symbolic of a herd animal (e.g., sheep) with observable signals ("tells"), and another (sheepdog) that infers behaviors and takes actions to achieve goals like barrier coverage or collision avoidance. Applications span pen testing, game theory, antiterrorism, and more. Scope challenges include variations in animal models, implementations, and goals. In a graduate class at MS&T, I experimented with reinforcement learning using a kindergarten hallway collision scenario, but struggled with incentives for emergent behavior, resulting in a B grade. Future work could focus on extracting governing dynamics for publishable insights.

A .NET-based deals engine service featuring Swagger for API documentation, Nginx as a reverse proxy, and Cloudflare for hosting. Supports dynamic configurations for production and local environments, with focus on scalability and security.

An HTML/JS application using Leaflet for interactive maps, integrating latitude/longitude data for cruise destinations like Port Charlotte, with features for scheduling and visualization.