I’m a computer science student who builds practical ML systems and the software around them.
Most of my work sits where data meets production:
- Time-series forecasting and feature engineering for noisy, real-world signals
- Cybersecurity analytics and lightweight detection for constrained/edge settings
- Serverless backends and automation that are simple to operate
I like projects that force clean interfaces and careful trade-offs (latency, cost, reliability). A few themes I’ve worked on:
- Hybrid intrusion detection for IoT-style traffic using a mix of rules/fuzzy logic, anomaly detection, and compact neural models
- Market forecasting experiments that combine OHLCV-style sequences with sentiment indicators
- A serverless comment moderation service with abuse resistance (verification, honeypots, rate limits) and async AI-based triage
- End-to-end tabular ML pipelines (preprocessing, model selection, and a small web UI for inference)
This site is mostly notes and write-ups on what I’m learning as I build: ML workflows, systems glue, and the occasional debugging story.
