Rakan Abu Awwad

Technical Analyst & Developer Intern

Virginia Tech, Greater Washington D.C. Area (GWDCA) · Alexandria, VA

• Led requirements analysis across multiple Virginia Tech departments to design a centralized asset management workflow, replacing manual Excel tracking for 60+ users
• Mapped and documented a seven-stage procurement-to-inventory process, defining data ownership and handoff points across four stakeholder groups
• Modeled the relational schema and configured role-based views, per-department pending-request dashboards, room/building filtering, and required vs. optional field rules to eliminate duplicate entry across three systems

Graduate Research Assistant

Virginia Tech, Dept. of Computer Science · Blacksburg, VA

• Fine-tuned and quantized TinyBERT (PyTorch, Hugging Face Transformers) for real-time PII and sensitivity classification, achieving 93.7% accuracy at sub-100ms on-device inference; paired transformer-based detection with a regex rules engine to reduce false negatives on structured identifiers
• Generated a 10,000+ synthetic email dataset using Faker and GPT-4 prompt engineering to overcome labeled-data scarcity, enabling supervised fine-tuning without exposure to real user data
• Designed a privacy-preserving inference architecture with zero external data transmission, serving the model locally via a Manifest V3 Chrome Extension and Flask backend so sensitive content never leaves the user’s device, addressing enterprise data-governance and GDPR constraints

Software Engineering Intern

Rubu Al-Quds Technology · Amman, Jordan

• Built and deployed 15+ RESTful API endpoints using FastAPI (Python), PostgreSQL, and Docker for inventory and order management microservices, handling 10,000+ daily requests across a 5,000+ product catalog
• Optimized PostgreSQL performance through targeted indexing, query refactoring, and SQL tuning, reducing query latency by 40%
• Implemented JWT authentication and role-based access control (RBAC) across 10+ secured endpoints using FastAPI Security and PyJWT, enforcing data integrity and reducing inventory discrepancies by 60%

Software Engineering Intern

Virginia Tech Marketplace · Remote

• Contributed to an open-source platform for students to post and review local places
• Built a responsive multi-page frontend using Next.js and React with optimized routing, reducing redundant data fetching and improving perceived load performance
• Developed a full-stack review system and JWT-based auth restricted to verified .edu emails via FastAPI and PostgreSQL, securing backend access and enabling verified user feedback

Car Dealership Management System

In productionamericanautosalesvirginia.com

• Architected and deployed a full-stack dealership platform (Next.js, Node.js, Express, MySQL) with dynamic inventory browsing, server-side rendering (SSR), and real-time search/filtering across 100+ listings, serving 2,500+ users
• Built a secure admin dashboard with JWT auth and role-based access control (RBAC) for end-to-end listing management, integrating Cloudinary and AWS S3 for media storage and deploying via Vercel with CI/CD

Next.js · Node.js · Express · MySQL · AWS S3 · Cloudinary · Vercel

Click anywhere or press Esc to close

Email Sensitivity Detector

Graduate researchDemo videoSource

• Chrome extension that reads Gmail drafts as you type and highlights names, addresses, card numbers, and other PII before you hit send
• Fine-tuned and quantized TinyBERT paired with a regex rules engine, served from a local Flask backend: 93.7% accuracy, under 100ms, and no email content ever leaves the device
• Trained on a 10,000+ email synthetic dataset generated with Faker and GPT-4, so no real user data was needed

PyTorch · Hugging Face · TinyBERT · Flask · Chrome Extension (MV3)

Play demo · 0:31

1. Gmail composeContent script reads the draft as you type
2. Local Flask APIListens on localhost only
3. TinyBERT + regex rulesQuantized model, under 100ms per pass
4. Inline highlightsFlagged spans marked in the draft

No external network calls. Email content never leaves the machine.

M.S. in Applied AI & Machine Learning

Virginia Tech · GPA 4.00 / 4.00

B.S. in Computer Science

Virginia Tech

Information Privacy and Security (IPS)

CITI Program · Credential ID 72212946