I build AI systems that work in the real world.
Austin-based builder, researcher, and technical PM. First-generation student at ACC. Shipping production AI while studying full-time.
I'm a full-stack AI developer and technical program manager based in Austin, TX. My work spans production RAG systems, real-time applications, and AI governance at the institutional level. I build things that solve real problems for real organizations, then measure whether they actually worked.
Current Work
Feb 2026–Present. Product Manager for EDGE Team (ACC's pilot entrepreneurship program): leading Nest Fest Spring and Fall 2026, sponsor pipeline, committee recruitment, and faculty outreach across 30+ professor meetings. Managing a cross-functional team of 8–10. Austin Future Makers: digital operations and content. Designing the email journey, maintaining program sites, managing content pipeline. Attends inter-institutional planning sessions led by ACC Head of Product Dr. Andrea Kehoe. Student Representative, ACC Collegewide AI Strategic Planning Committee (CAISPC).
Jun 2021–Present. Full-time IT security and incident response for 200+ law firm clients, 40 hours per week while attending college full-time. Bilingual support in English and Spanish.
Austin Community College | 2024–Present. Lead a 45+ member student innovation organization. Product strategy and roadmap for Nest Fest 2025 (completed) and dual 2026 editions. Faculty strategic planning representative.
Featured Projects
Small nonprofits and community organizations spend 20–40 hours manually searching for grant funding they qualify for. Most never find it. I built a Graph RAG system that does that search in half a second.
Architecture: hybrid knowledge graph (Neo4j) + vector similarity search (Pinecone) + local LLM inference (Ollama) + FastAPI backend + Next.js frontend. 85–95% relevance score in production.
I designed a controlled benchmarking study across Gemini, Claude 3.5, and ChatGPT-4o, building the same product on each platform to isolate performance variables and measure real capability differences under business constraints.
Output: a solution recommendation framework mapping each model's strengths to specific use cases by business owner type. Presented findings to 40+ attendees at SXSW EDU (March 2026) and 20+ at ACC's AI Power Lunch (March 2026).
This is not a benchmark. It is a research methodology for organizations deciding which AI platform to build on.
Currently migrating from Vercel to Cloudflare Workers and Durable Objects, researching stateful edge computing patterns for real-time applications where latency is a UX problem, not just a performance metric.
Refactored and stabilized a HIPAA-adjacent healthcare benefits optimization platform. Untangled a CSS and JavaScript dependency web that made simple changes break unpredictably, rebuilt with self-contained modular architecture.
The lesson: complexity is a product problem before it is a code problem.
In Progress
Psychological support system for long-duration spaceflight. Built for RiverHacks. Analyzes emotional state across multiple dimensions using vector memory. Ongoing development, exploring applications in high-isolation environments. Stack: Emotion AI, Vector Memory.
Design concept developed after observing a communication breakdown between front desk and triage. Prototyped a dual-dashboard system connecting labs, front desk, and patient rooms in real time. Not a production deployment. A design exercise and proof of concept.
Recognition & Speaking
"Business in a Box: AI and Entrepreneurship Lab" | March 11, 2026 | Austin, TX | 40+ attendees. Co-facilitated alongside ACC Head of Product and Executive Vice Chancellor.
LLM Benchmarking Research | March 2, 2026 | ACC Make It Center | 20+ attendees.
"Determining the Roles Required for Effective AI Governance in Higher Ed." Featured as student representative on ACC's Collegewide AI Strategic Planning Committee.
"students can model inclusive innovation." LinkedIn, 2026.
Philosophy
I believe the highest-value thing AI can do is not replace human judgment. It is to put better information in front of human judgment faster.
The Grant Search Tool does not decide who gets funding. It makes the 205,000-option search problem solvable for a two-person nonprofit team in 30 seconds instead of 30 hours. That is the design philosophy behind everything I build.
The ROI is not in creating systems that remove the human element. It is in creating agentic systems that allow humans to achieve more and produce better output.
Open to Summer 2026 Internships
Looking for opportunities in AI engineering, product management, and security. Based in Austin, TX.
mailabel.rincon@g.austincc.edu