At FlyRank AI, I built a RAG system with LlamaIndex, including semantic chunking, a query-rewriting retriever, and an LLM API with retries. I’m completing a mentor-reviewed backend track focused on API contracts, structured outputs, and failure handling in AI services.
On Eskom Grid Cloud, I provisioned a serverless AWS platform with Terraform using EventBridge, Step Functions, Lambda, S3, and ECR. I also applied least-privilege IAM and set up GitHub Actions to plan every pull request through OIDC, with no stored keys.
I’m building Qonda, an AI tutor using ASP.NET Core and Blazor WebAssembly, designed around PostgreSQL and swappable LLMs. Changes go through pull requests with CI and xUnit tests.
Across projects, I’ve developed applications including a live Next.js app connected to MongoDB Atlas, a React and TypeScript agricultural data platform, and a Python/FastAPI service with real-time WebSocket communication. At WeThinkCode_, I mentor students in software engineering and application development.

