DEMARICK WEBB-RIVERAFULL STACK DEVELOPER / SYSTEMS ADMINISTRATOR / U. S. ARMY VETERAN / ACTIVE CLEARANCE

OVER 200,000 SERVICE MEMBERS TRANSITION OUT EVERY YEAR, AND 44% FEEL UNPREPARED. EXISTING RESOURCES ARE SCATTERED, GENERIC, AND HARD TO NAVIGATE. VETERANS DESERVE BETTER.

I BUILT A FULL STACK PLATFORM USING NEXT.JS AND GPT-4O TO CREATE PERSONALIZED ROADMAPS. I INTERVIEWED 12 VETERANS TO UNDERSTAND REAL PAIN POINTS. MULTI-STEP ONBOARDING WITH ZUSTAND STATE MANAGEMENT AND STRUCTURED PROMPTS ENSURE EVERY USER GETS A TAILORED PLAN, NOT A GENERIC CHECKLIST.

PRODUCTION-DEPLOYED PLATFORM WITH A 95% TASK COMPLETION RATE IN USER TESTING. BUILT WITH ZERO-TRUST SECURITY PRINCIPLES TO PROTECT SENSITIVE PII. THIS IS THE TOOL I WISH I HAD WHEN I TRANSITIONED.

NASHVILLE TRAFFIC IS NOTORIOUS, BUT EXISTING PREDICTION TOOLS ARE EITHER INACCURATE OR BUILT FOR CITY PLANNERS, NOT COMMUTERS. PEOPLE NEED ACTIONABLE INSIGHTS, NOT RAW DATA DUMPS.

I BUILT AN ML-POWERED DASHBOARD WITH A REACT FRONTEND AND PYTHON BACKEND USING SCIKIT-LEARN. TRAINED MODELS ON HISTORICAL TRAFFIC DATA TO PREDICT CONGESTION PATTERNS. POSTGRESQL FOR DATA PERSISTENCE. FOCUSED ON MAKING COMPLEX ML OUTPUTS ACTUALLY USABLE FOR REGULAR PEOPLE.

15% BETTER PREDICTION ACCURACY THAN THE CITY'S EXISTING TOOLS. EARNED AN A IN MY MTSU CAPSTONE COURSE. PROVED I CAN TAKE A PROJECT FROM DATA COLLECTION TO DEPLOYED PRODUCT.

PHISHING ATTACKS ARE GETTING MORE SOPHISTICATED, AND NON-TECHNICAL USERS ARE THE MOST VULNERABLE. MOST PEOPLE CANNOT SPOT A MALICIOUS LINK UNTIL IT IS TOO LATE.

BUILT A SECURITY-FOCUSED TOOL USING NEXT.JS THAT INTEGRATES MULTIPLE THREAT INTELLIGENCE APIS. REAL-TIME URL ANALYSIS WITH CLEAR, ACTIONABLE RESULTS. NO SECURITY JARGON — JUST A SIMPLE VERDICT WITH CONTEXT.

REAL-TIME THREAT DETECTION WITH SUB-SECOND RESPONSE TIMES. SECURITY-FIRST DEVELOPMENT PRACTICES INCLUDING INPUT SANITIZATION AND RATE LIMITING. A TOOL THAT ACTUALLY PROTECTS PEOPLE.

VANDERBILT'S AUTONOMOUS VEHICLE RESEARCH TEAM NEEDED A WAY TO STUDY HOW DRIVERS REACT WHEN AUTONOMOUS SYSTEMS HAND BACK CONTROL. COGNITIVE LOAD AND REACTION TIME ARE HARD TO MEASURE IN THE REAL WORLD.

BUILT A UNITY-BASED DRIVING SIMULATOR WITH SPLIT-SCREEN COGNITIVE MINIGAMES TO TEST ATTENTION UNDER LOAD. C# LOGIC FOR REAL-TIME DATA COLLECTION AND SPEECH RECOGNITION FOR VOICE COMMANDS. DESIGNED TO GENERATE RESEARCH-GRADE DATA.

RESEARCH-GRADE SIMULATOR DEPLOYED FOR ACTIVE STUDIES. ENABLES REAL-TIME ATTENTION MONITORING AND SAFER AUTONOMOUS TRANSITIONS. WORKING WITH PHD RESEARCHERS TAUGHT ME A LOT ABOUT BUILDING FOR EXACTING REQUIREMENTS.