Conduct a comprehensive qualitative study to understand user awareness and behaviors regarding cybersecurity threats on social media platforms, specifically TikTok and Facebook.

This research project involved interviewing over 100 active TikTok and Facebook users to assess their understanding of social media security threats, privacy concerns, and protective behaviors. The study aimed to identify knowledge gaps and develop actionable recommendations for improving cyber-awareness among social media users.

Social media platforms have become primary targets for cybercriminals, yet user awareness of security threats remains critically low. Many users unknowingly engage in risky behaviors that expose them to phishing, identity theft, account takeovers, and other cyber attacks.

Lead Researcher & Cybersecurity Analyst — Designed the research methodology, conducted user interviews, analyzed qualitative data, identified security awareness patterns, and developed comprehensive recommendations for users and platform providers.

• Semi-structured interviews with 100+ TikTok and Facebook users
• Qualitative data analysis using thematic coding
• User behavior observation and pattern identification
• Threat landscape analysis across social media platforms
• Literature review of existing social media security research
• Cross-platform security comparison studies

• Phishing and social engineering attacks on social media
• Privacy settings awareness and configuration
• Account security practices (passwords, 2FA, recovery options)
• Recognition of fake accounts and impersonation attempts
• Malicious link identification and safe browsing habits
• Data sharing behaviors and privacy implications
• Awareness of platform-specific security features
• Response to suspicious messages and friend requests

• Identified critical gaps in social media security awareness among everyday users
• Developed evidence-based recommendations adopted by university IT security programs
• Created user-friendly security guidelines distributed to 500+ social media users
• Informed the development of targeted security awareness campaigns
• Contributed to academic understanding of social media threat perceptions
• Provided actionable insights for improving platform security education
• Highlighted the need for better default security settings on social platforms

Replace a paper-and-pen queue workflow with a digital system that automates customer notifications and gives the counter a live view of who's next — so staff spend less time texting and more time serving customers.

A single-page web app the shop runs in the browser. Staff check customers in with name and phone, the queue updates live, a one-tap "Ready" button fires off a WhatsApp/SMS notification, and a second screen facing the counter displays the current queue for walk-in customers to see their position at a glance.

The shop was tracking orders on paper clipboards and texting each customer individually from a staff phone. This led to missed pickups, mis-ordered notifications, and lost slips during busy hours. There was no visibility for customers on where they stood in line.

Full-stack developer — gathered requirements on-site with the shop owner, designed the dual-screen UX (staff panel + counter display), built the app as a single-file HTML deliverable, deployed it to Cloudflare Pages, and trained staff on day one.

• localStorage for persistence — survives tab reloads and works fully offline
• WhatsApp click-to-chat deep links for notifications (no paid SMS gateway needed)
• Dual-view architecture: staff admin panel + counter-facing display screen
• Phone number normalization and validation to prevent mis-sends
• Daily queue reset with served-customer history retained for metrics
• Deployed as single-file HTML for zero-config hosting on Cloudflare Pages

• Eliminated paper queue lists and one-by-one manual texting
• Cut customer pickup confusion and staff notification overhead significantly
• Deployed to production and in active daily use at a Philadelphia butcher shop
• Zero recurring cost — no SMS gateway subscription, no backend servers
• Trained non-technical staff to use and reset the system themselves

Help job seekers actually make it past ATS filters — by turning vague feedback like "tailor your resume" into a concrete, AI-generated diff they can apply in minutes.

A browser-based application where the user uploads a resume (PDF or DOCX) and pastes a job description. The Claude API analyzes both, returns an ATS-compatibility score with an explainable breakdown, flags missing keywords, and generates rewritten versions of weak bullet points using metric-driven language. The final report exports as a clean, branded PDF.

Most resumes never reach a human — ATS systems screen them out based on keyword and format mismatches before a recruiter ever sees them. Candidates know this, but existing tools either charge monthly subscriptions, give generic advice, or don't actually rewrite content. There's no fast, private, affordable way to get specific fixes.

Solo product developer — end-to-end ownership of architecture, Claude API integration, prompt engineering, UX design, PDF generation, demo mode, and go-to-market strategy (Product Hunt launch planned).

• Claude API for resume analysis, scoring, and bullet rewriting
• Mammoth.js to parse DOCX files client-side without a server round-trip
• jsPDF for branded PDF report export with structured sections
• Demo mode with sample resumes and canned responses — users can try it without an API key
• Privacy-first: all processing is client-side; resume content is never stored
• Shipped as a single-file HTML for easy distribution and self-hosting

• ATS compatibility score with an explainable, section-by-section breakdown
• Keyword gap analysis — shows exactly which job-description terms are missing
• AI-powered bullet rewriting — weak bullets become metric-driven, action-oriented
• Polished PDF report export with before/after comparisons
• Demo mode for users without a Claude API key
• Works offline-first after initial load

• Product Hunt launch in preparation as a monetizable standalone tool
• Proof-of-concept for shipping Claude API in a single-file, client-side product
• Privacy-preserving by design — no servers, no data retention
• Zero hosting cost — deployable to any static host

Build a secure, reproducible AWS network foundation as code — so every future cybersecurity project (honeypot, threat detector, etc.) starts from the same hardened baseline and can be torn down and rebuilt with a single command.

A Terraform module that provisions a production-style AWS network: a VPC across two Availability Zones with public and private subnets, a bastion host for secure admin access, a private EC2 instance, a NAT Gateway for outbound traffic, an encrypted S3 bucket for honeypot logs, an SNS topic for alerts, and IAM roles scaffolded for a future serverless threat detector reading CloudTrail events.

Clicking through the AWS console to set up secure networking is error-prone, non-reproducible, and hard to audit. Every cybersecurity project I wanted to ship (honeypot, SIEM feeds, threat detection) needed the same network foundation — and manually recreating it each time was wasteful and introduced configuration drift.

Infrastructure architect and IaC developer — designed the network topology, wrote the Terraform modules, configured security groups and IAM with least-privilege defaults, and validated the full stack through terraform plan / apply / destroy cycles.

• Modular Terraform design — network, compute, storage, and IAM kept in separate modules
• Multi-AZ deployment for high availability (2 AZs, public + private subnets each)
• Least-privilege IAM roles scoped to specific services and actions
• Encrypted S3 bucket with versioning enabled for honeypot log retention
• SSM Session Manager option on bastion for keyless admin access
• SNS topic and IAM roles pre-scaffolded for a future Lambda-based threat detector
• Sensitive values (key pair names, alert email) parameterized via tfvars

• VPC spanning two Availability Zones with public and private subnets
• Bastion host in public subnet with SSH and SSM Session Manager access
• Private EC2 instance reachable only via the bastion
• NAT Gateway providing outbound internet for the private subnet
• Encrypted, versioned S3 bucket dedicated to honeypot log ingestion
• SNS alert topic ready for security notifications
• IAM roles scaffolded for a future serverless CloudTrail threat detector

• Production-ready network foundation shared across three cybersecurity projects
• Demonstrates IaC best practices — least privilege, encryption at rest, multi-AZ
• Full stack deployable or destroyable with a single terraform apply / destroy
• Reusable baseline that eliminates manual console drift between environments
• Serves as a portfolio artifact demonstrating cloud security engineering