I'm a Software Engineer from Zimbabwe, based in the US — a CS grad (BCU, 4.0 GPA, May 2026) with three internships at Meta spanning full-stack web, backend infrastructure, and cross-platform mobile. I care a lot about building things that actually work for people, and I'm always thinking about the gap between "technically correct" and "genuinely useful."
Meta — 3x Software Engineering Intern (2023 · 2024 · 2025)
Three summers, three completely different problem spaces:
- 2025 (Mobile, Cross-Platform): Built a Facebook feature for searching and bulk-adding saved content to collections. Shipped 15+ reusable React Native components that saved 300+ engineering hours and cut the dev-to-launch timeline by 30%.
- 2024 (Backend Infrastructure): Designed a multi-tier FAQ system that improved chatbot lead retention by 10% and cut business page response time by an estimated 40% using a tree-based GraphQL schema.
- 2023 (Full-Stack Web): Delivered a full-stack web prototype in 5 weeks with zero prior JavaScript experience. That summer changed everything.
Basafy — AI Job Search & Interview Tracker
A React Native mobile app that parses Gmail using Python and Google APIs to automatically update a real-time dashboard of job applications, recruiter contacts, and interview stages. Integrated calendar APIs and custom logic to surface AI-driven insights, set deadline reminders, and track where every application stands — built on a full-stack workflow using React Native, Node.js, and Supabase.
Built it because tracking 100+ applications in a spreadsheet was genuinely unsustainable, and I wanted something that worked the way my brain does.
Flux — VS Code Extension for Structured GitHub Collaboration
Flux started from frustration. After working with internal engineering tooling at Meta, returning to traditional GitHub pull request workflows felt surprisingly primitive — giant PRs, endless browser tabs, context switching, and code reviews that felt more like administrative overhead than actual collaboration.
Flux is a VS Code extension that adds a structured collaboration layer directly on top of GitHub. It reimagines version control and code review as visual, layered, and context-rich — without ever leaving the editor. Developers can work in smaller, self-contained changesets, attach screenshots and demo recordings, annotate architectural decisions, visualize dependencies between changes, and track collaborative review tasks, all in one place.
The larger vision is a smarter collaboration operating system for software teams: combining version control, review systems, AI-assisted engineering, and visual workflow management into one unified experience — built for everyday developers, not just those with enterprise infrastructure.
A real-time AI-driven makeup visualization tool built with Python, OpenCV, and Mediapipe facial landmarking — allowing users to try on 20+ makeup styles through their camera without touching a single product.
The technical foundation uses Mediapipe's facial landmark detection to map precise regions of the face in real time, then applies CycleGAN image style transfer to render makeup overlays that adapt naturally to the user's features, lighting, and skin tone. The pipeline runs entirely in Python, with OpenCV handling the video stream and rendering layer.
What made this project personal was the dataset problem. Most existing computer vision datasets for makeup and facial recognition are heavily skewed toward lighter skin tones, which means applications built on them perform noticeably worse for people with darker complexions. I curated a 1,000+ image dataset specifically designed to represent diverse skin tones with professional makeup application — not as an afterthought, but as the foundation. The result was a 30% improvement in application accuracy, and more importantly, a tool that actually works for the people who have historically been left out of these systems. Building for diversity is not a feature. It is an engineering decision, and it changes the product.
- 🥉 IEEE SoutheastCon Hardware Competition (2025) — Developed autonomous movement logic for a robotics challenge and assembled the robot chassis from scratch. Placed 3rd out of 60 schools.
- Jane Street INSIGHT Program — Built a Python trading bot in a simulated competition. Placed 3rd overall.
- Generation Google Scholarship (2024) — Selected as 1 of 56 recipients across North America for leadership potential and commitment to diversity in tech.
- Black Excellence SMART Hackathon — American Airlines (2024) — Built a digital assistant web prototype in under 24 hours with a team of 4. Placed 4th out of 47 schools.
I play chess seriously — the kind of seriously where you're analyzing your own games at midnight trying to figure out where you went wrong. There's something about chess that scratches the same itch as engineering: pattern recognition, planning multiple steps ahead, and knowing when to be patient versus when to commit. I've been improving steadily and I intend to keep it that way.
On the robotics side, I'm an active member of BCU's IEEE Robotics Club, where hardware meets software in ways that are genuinely humbling. The IEEE SoutheastCon competition was one of the most hands-on engineering experiences I've had outside of Meta — building autonomous movement logic for a robot we assembled from scratch, competing against teams from 60 universities, and placing 3rd. It reminded me that the best engineering happens when you're forced to make something work in the real world, not just in a pull request.
I grew up in Zimbabwe and ended up in Florida studying CS at an HBCU. That path has shaped how I think about access, representation, and who technology actually gets built for — it is why I care so much about inclusive datasets, software that considers everyone, and being part of communities like NSBE and IEEE where those conversations happen naturally.
I spend probably too much time thinking about why things are the way they are. That curiosity is what keeps me building.