Home
/Home_
I am Riz.
I work on machine learning, software, and the cloud infrastructire for both.
MS in Computer Science from SEMO. Finishing up an MS in AI at UT Austin.
This site is where I put what I learn. Visual guides for the math that comes up over and over in ML.
Notes on the systems that sit underneath ML in production. A few research papers from graduate work.
If you want the latest, the Blog is the dated stream.
Featured¶
- Bayes' Theorem, visualized. Prior, likelihood, posterior, and the intuition traps that bite even after you've done the math.
- Information Theory, visualized. Entropy, cross-entropy, KL divergence, mutual information. Why cross-entropy is the loss function it is.
- Sketch Classification with CNNs. Deep learning on 250 classes of hand-drawn sketches.
- Senator Tweet Analysis. NLP over 419K+ tweets from the 117th Congress.
- LRU Cache. O(1) get/put with a hash map and a doubly linked list.
- Queue-Based Stage Decoupling. What to do when one stage is faster than the next.
ML and DS¶
- Probability Basics. The probability scale, conditional probability, combining events.
- Compound Probability. Sequential events, independence, expected value.
- Random Variables and Distributions. Discrete vs. continuous, PMF, PDF, CDF.
- Statistics Fundamentals. Descriptive vs. inferential, center and spread, correlation.
- Bayes' Theorem. Prior, likelihood, posterior, sequential updating.
- Information Theory. Entropy, cross-entropy, KL divergence, mutual information.
Publications¶
- Sketch Classification with CNNs. 250-class hand-drawn sketch recognition with deep learning.
- Image Recognition with Classical ML. HOG + SVM facial recognition without deep learning.
- Senator Tweet Analysis. NLP-driven content analysis of 419K+ tweets.
- English to ASL Translator. Web app for ASL translation with a usability study.
- Cloud Asset Management for Hospitals. Cloud adoption analysis under HIPAA.
- Vulnerability Scanning of Web Applications. OWASP scanning across 50 top sites.
Systems¶
- Queue-Based Stage Decoupling. High-throughput message processing architecture.
- Ingest-Transform-Serve Pipeline. Scheduled ingestion to a REST API.
- Memory Allocation and Fragmentation. The malloc/free patterns under GPU caching allocators.
- Session Window Feature Engineering. Sessionization for recommendation and churn models.
- All 30+ posts on algorithms, data structures, and microservices.