High-Performance Systems & Infrastructure Engineer
Specializing in HPC, Cloud-Native AI, and Autonomous Systems.
I am a Computer Engineering student at Queen's University (Class of 2026) with a deep focus on monumental open-source architectures and sustainable AI infrastructure. My work bridges the gap between bare-metal hardware constraints and distributed multi-host orchestration, from optimizing Solana's blockchain infrastructure to architecting carbon-aware LLM serving on Kubernetes.
- Solana Labs (Fellow): Engineering high-throughput blockchain infrastructure and applied ML.
- vLLM / llm-d: Architected an Energy-Aware Endpoint Picker (EPP) in Go, integrating eBPF telemetry and Carbon-Aware Controllers to reduce inference energy by 17.4%.
- Kubernetes (SIG Apps): Core contributor to LeaderWorkerSet (LWS); implementing state-machine fixes for distributed GPU scheduling and high-availability pod orchestration.
- Deel (ML Research): Built a scalable legal-data RAG pipeline using Python/Selenium/Pinecone, automating ingestion of 10,000+ documents with structure-aware semantic chunking.
- Queen's AutoDrive (CAN Lead): Architected a multi-threaded ROS2 C++ node for SAE Level 4 vehicles. Managed 100+ signals across 4 isolated CAN networks with <2ms latency and ISO 26262 ASIL-D compliance.
- Foresoon (Robotics Intern): Deployed a CV pipeline (OpenCV/ArUco) for a Kinova Gen3 6-DOF arm on a Clearpath Jackal UGV for autonomous assistive feeding.
- G.O.S. Phytotron (IoT): Engineered a custom 4-layer PCB (nRF52840) running Zephyr RTOS with Thread mesh networking, achieving 360nA quiescent current.
- HPC Capstone (ELEC498): Developed a PID-controlled DVFS engine for miniMD, utilizing Intel RAPL and Beta-Adaptation algorithms to optimize MPI communication phases.
- Arista Networks (Intern): Modeled $2M project portfolios in SAP S/4HANA, flagging $150K in overruns and building Power BI dashboards to boost project efficiency by 25%.
- Queen's HPC (Co-Founder): Established the first ratified HPC club at Queen's; developed a technical curriculum covering SLURM, MPI/OpenMP, and HPL/HPCG benchmarking.
| Category | Tools & Technologies |
|---|---|
| Languages |          |
| Infra & Cloud |         |
| Systems & RTOS |        |
| AI & Data |        |
| HPC & Net |      |
A phase-aware Kubernetes scheduling plugin for the llm-d framework. It implements an adaptive, closed-loop controller to route LLM Prefill and Decode phases across heterogeneous accelerators (GPUs and ASICs) to optimize for Software Carbon Intensity (SCI) and energy budgets.
Phase-aware Dynamic Voltage and Frequency Scaling (DVFS) controller for the miniMD proxy application.
🚀 ELEC-475
Labs and classwork from ELEC 375
Level 4 autonomous vehicle real-time road-sign TensorFlow Lite object detection model trained and deployed on Coral TPU edge hardware using EfficientDet-Lite.
Generated with ⚡ by Manus | Last Updated: 2026-06-13 04:07:40 UTC