Projects
Systems built, AI integrated, and lessons learned in production
Personal Projects
AI-Powered Interactive Resume (this site)
Built this site as a working demonstration of AI integration into a practical system. The core is a RAG (Retrieval-Augmented Generation) pipeline: structured data about my background is embedded using OpenAI's embedding models, and user questions are matched against the most semantically relevant context before being passed to an LLM for a conversational response.
Architecture decisions: Separated knowledge into domain-specific JSON files (personality, career goals, behavioral examples, etc.) so retrieval stays focused. The chunking strategy and cosine similarity threshold were tuned to balance precision against recall. Deployed as a serverless API on Vercel.
Why it matters: This is exactly the kind of work I want to do: identifying a real problem (a static resume tells you nothing about how someone thinks), designing an architecture to solve it, and implementing AI as the mechanism rather than the product. The AI serves the use case; it isn't the use case.
News Aggregation & User Platform
Full-stack platform pulling from multiple global news sources, categorizing articles into storyline threads, and surfacing them through a user-controlled interface with accounts, real-time WebSocket chat, and a dashboard. Built to reduce single-outlet bias in news consumption.
Architecture owned end-to-end: Database schema, user authentication flow, background aggregator, WebSocket real-time layer, and frontend. Every architectural decision (how the aggregator feeds the database, how sessions are managed, how data flows from API to UI) was mine to make and get right.
Hardest part: Security architecture. Authentication systems are easy to get wrong in ways that look correct until they aren't. Getting session tokens, password hashing, and route protection right required understanding the threat model, not just following a tutorial.
AI Workflow Automation Experiments
Ongoing series of experiments integrating AI tooling into real workflows: automating repetitive tasks, building LLM-assisted scripts, and evaluating where AI genuinely reduces friction versus where it introduces new complexity.
Focus: Practical implementation over theoretical understanding. What does it actually take to get an AI tool working reliably inside an existing process? Projects include prompt engineering for structured output, AI-assisted documentation generation, and lightweight agentic task runners.
Key insight: AI integration fails most often not because the model is wrong, but because the surrounding system wasn't designed to handle model uncertainty. Robust AI implementation requires thinking about the whole system, not just the model call.
Class Repositories
LIS 4376 - AI Applications
Applied AI coursework covering ML model evaluation, NLP, AI ethics, and real-world deployment considerations. The emphasis was on critical evaluation: understanding when AI outputs should be trusted, when they shouldn't, and what that means for system design.
Key takeaway: Developed the ability to assess AI tools through both a technical and ethical lens. I can evaluate where AI creates genuine efficiency versus where it introduces bias, risk, or unnecessary complexity: which is exactly the analysis a systems consultant needs to do before recommending an AI integration.
View RepositoryLIS 4381 - Mobile Web Development
Responsive, mobile-first web applications with focus on system architecture, CSS layout systems, and cross-device performance. Designing for multiple screen sizes simultaneously requires making explicit tradeoffs: what gets priority, what gets cut.
Key takeaway: Small screens force ruthless prioritization. That discipline of stripping a system down to what actually earns its place is directly transferable to consulting and architecture work, where scope creep is the default failure mode.
View RepositoryGitHub Profile
Personal experiments, coursework, and ongoing development projects spanning web development, AI integration, automation, and systems design. The history shows how skills have developed over time: including works in progress, not just finished products.
Visit GitHub