Curriculum vitae
Super Washington Banga
Software Engineer · Backend Systems · Cloud Platforms · Applied AI
Computer Science undergraduate · University of Zimbabwe
- Harare, Zimbabwe
- superwbanga@gmail.com
- github.com/swbanga
- linkedin.com/in/swbanga
Summary
Computer Science student and software engineer building backend, cloud, and applied AI systems. Work includes an immutable .NET ledger, a FastAPI healthcare intake pipeline with field-level encryption, and GradKit’s queued AI document workflow on Azure Container Apps. I focus on typed interfaces, testable domain logic, infrastructure as code, and controlled model integration.
Selected Engineering Work
Ledger-Core v1.0 · .NET 10 financial systems
2026 — Present- Implemented an append-only double-entry ledger in C# and .NET 10 using Clean Architecture, DDD aggregates, and CQRS with MediatR.
- Moved event delivery to a transactional T-SQL outbox; Redis idempotency makes at-least-once RabbitMQ delivery safe to replay.
- Added RowVersion optimistic concurrency with bounded retries and a 26-test containerized integration gate.
Med-Core v1.0 · FastAPI healthcare intake
2026 — Present- Built an async FastAPI and SQLAlchemy intake path that validates, authorizes, and encrypts protected fields before PostgreSQL persistence.
- Implemented scoped OAuth2/JWT access, Redis-backed revocation, rate limiting, idempotency, and immutable access auditing.
- Defined the Azure Container Apps deployment in Terraform and added dependency and Trivy image scanning to GitHub Actions.
GradKit Platform · AI document processing SaaS
2026 — Present- Moved DeepSeek document generation from synchronous requests to Redis-backed jobs so model latency does not block the web path.
- Put the Node.js application behind an Nginx edge with rate limits, CSP, and HSTS on Azure Container Apps.
- Replaced on-disk PDF staging with transient Playwright rendering so completed documents leave no residual server copy.
AI Systems Roadmap
Model capabilities sit behind the same schemas, queues, authorization, and observability used for other production workloads.
Applied AI Engineering · Active
- Structured outputs with Pydantic and Zod contracts
- Queued model jobs and guarded execution boundaries
- Next: pgvector RAG, deterministic tools, and provider failover
Mathematics & Classical ML · Foundation
- Linear algebra, calculus, probability, and statistics
- NumPy, Pandas, scikit-learn, and time-series work
- Progression: PyTorch → Transformers → LoRA fine-tuning
Education & Certifications
BSc Computer Science · University of Zimbabwe · Undergraduate