Software Engineering Student @ McMaster University
I'm passionate about building robust, scalable software solutions with a strong focus on high-performance backend architecture, machine learning infrastructure, and GPU-accelerated computing. I enjoy turning complex problems into highly optimized, automated, and agentic systems.
| Area | Technologies |
|---|---|
| Languages & Compute | |
| AI & Machine Learning | |
| Cloud & Architecture | |
| Data & Databases | |
| Frameworks & APIs |
| Project | Description | Tech Stack |
|---|---|---|
| OmniDoc: Multimodal Document Intelligence | Built an inference pipeline and batching logic running Llama 3.2 Vision and Qwen-VL simultaneously on AMD MI300X hardware. Achieved 340 pages/min (18x faster than CPU baseline) for complex chart-level Q&A and semantic citations. | Python ROCm Llama Vision Qwen-VL MI300X |
| ML Experiment Tracker | Developed a high-throughput polling dashboard for tracking ML training runs. Resolved read-heavy bottlenecks by placing Redis in front of PostgreSQL with write-invalidations, achieving sub-100ms reads under load. | React Flask Redis PostgreSQL Docker |
| CUDA Matrix Multiplication Engine | Engineered a low-level GPU compute kernel utilizing shared memory tiling and memory coalescing. Profiled heavily with NVIDIA Nsight Compute to resolve compute-bound vs memory-bound bottlenecks on massive matrix workloads. | CUDA C++ NVIDIA Nsight GPU Profiling |
| PathFinderAI: Agentic RAG System | Architected a multi-agent RAG system with persistent state management across conversational turns. Integrated live external data APIs and semantic vector search using LangGraph for complex reasoning chains. | Python LangGraph LangChain Azure Vector DBs |