Forma Health allows patients to connect with doctors in a clinical study, using AI to allow easy logging and analytics of patients' journey through the study. I delivered the following components:

• Android App - Enables patients to collect and communicate specific information requested by a clinical study.

• Physician Portal - Website for physicians to configure and view study data.

• OpenAI LLM Processing - Developed system for LLM Prompt testing and optimization.

Developed multiple contributions to ARM Compute Library now included in production.

• GELU Activation Function - Added new activation function in both OpenCL and NEON. This required implementation of erf (error function) using NEON intrinsics.

• MeanStdDevNorm - Added int8 quantized implementation of normalization function. Added special case for F16 precision to avoid overflow issues.

• CLSubTensor - Fixed/Optimized map operation which previously called clEnqueueMapBuffer unnecessarily.

All changes were reviewed/optimized by myself and other ARMCL developers.

https://review.mlplatform.org/q/owner:%2522murray.kornelsen%2540mail.mcgill.ca%2522

Research: Optimizing neural networks, especially transformer models (BERT) on edge hardware through heterogeneous computing (CPU+GPU) and quantization.

Thesis: Low-Latency BERT Inference for Heterogeneous Multi-Processor Edge Devices

• Research focused on accelerating edge inference through heterogeneous computing.

• Developed a genetic algorithm method for optimizing the assignment of neural network operations to CPU/GPU in edge SoC. Used HiKey970 development board for testing.

• Combined heterogeneous computing with quantization. Developed an algorithm to optimize quantization configuration for pareto-optimal accuracy and latency.

• Developed an ARM Compute Library implementation of BERT for latency measurements.

COMP 310 - Operating Systems (Winter 2022 & Winter 2023)

• Presented tutorials on C Programming fundamentals: Introductory C Programming, Structs, Pthreads.

• Answered questions about course material through Ed forums.

ECSE 324 - Computer Organization (Winter 2022 & Winter 2023)

• Ran labs focused on ARM (32-bit) assembly programming.

• Answered student questions about ARM assembly.

• Graded student demos at end of each lab.

Worked as part of the software engineering team on production code to resolve bugs and customer issues. Worked on both backend (Java) and frontend (Angular) issues.

Wrote code to improve database analysis performance by taking advantage of database partitioning and parallelism. Researched and implemented for MySQL, Postgres, OracleDB, and MSSQL.