Recent engineering projects
View All
Tenant Payment System
Full-stack system for landlord-tenant workflows with payments, roles, and support tickets.
Study Companion
Adaptive study agent that reflects, prioritizes weak areas, and rebalances plans dynamically.
Enterprise Load Balancer
High-performance Layer 7 load balancer with circuit breakers, Redis caching, and Kubernetes deployment.
Class-Life Optimizer
Planning tool to help students balance work, classes, and recovery time effectively.
Burnout Prediction
ML research analyzing work-study student burnout patterns using TACC Stampede3 HPC.
Hello, I'm Dumisani
I'm a full-stack developer and ML researcher. My journey started with solving real problems — managing 50+ IT tickets weekly for 1,600+ users, optimizing MySQL queries to cut backend latency by 40%, and building AI agents that help people balance their work and personal lives.
What drives me is the intersection of intelligent systems and human-centered design. I don't just build software — I build agents that adapt, learn, and genuinely help people accomplish their goals. From ML models on TACC supercomputers achieving 88.6% accuracy to enterprise load balancers with 8 routing algorithms, I focus on turning insights into production-ready tools.
Presenting research at Greater Columbus Convention Center in Columbus, Ohio
My Philosophy
I design systems as autonomous agents that can reason, adapt, and take action — not just respond to inputs.
My projects stem from real research. The Class-Life Balance Optimizer was born from my PEARC25 burnout prediction study.
AI should amplify human capability, not replace it. I build tools that empower users to make better decisions.
From ML model training on HPC clusters to deploying production APIs — I own the full stack.