Gradient Descent vs Ordinary Least Squares
Working through linear regression from two angles: the iterative updates of gradient descent and the normal equation for ordinary least squares, with derivations and notes on when to use each.
Read articleMy path has taken me from mathematics and statistics to energy forecasting and production ML. At Zoom, I helped build anomaly detection and root cause analysis systems.
These days, I’m exploring AI agents, memory, and reinforcement learning — and keeping notes as I go.
Education
M.S. Computer Science · Georgia Tech
M.S. Probability & Statistics · Carleton
Experience
Machine Learning Engineer · Hangzhou
Built time series anomaly detection across 80+ services and 500+ monitors, alongside work on root cause analysis and agent workflows.
Data Scientist · Ottawa
Turned wind, solar, and energy demand data into forecasts for real-world energy decisions.
Research Associate · Ottawa
Applied data science to forecasting, computer vision, and recommendation systems.
BI & Reporting Analyst II · Ottawa
Made reporting more useful — and reduced reporting time by over 30% through automation.
Reinforcement learning meets sustainable land use in the Lake Malawi Basin.
Your tennis game, remembered.
Training notes and video become a lasting memory of your game. Rally connects patterns across sessions and builds a personal briefing for the next time you step on court.
Qwen Cloud Hackathon · MemoryAgent track
Designed and built as a solo project.
My AI learning notes, connected.
Notes from Algorithm, Engineering, ML Course, and Alisa’s LLM notebook. Explore their connections and read the notes.