Automated product matching: using ML, NLP to solve the entity-matching problem (noun-coreference problem).• Machine-Learning Modeling (1y 5m): Classifiers for product matching: text-matching classifiers, meta-model classifiers combining text, taxonomy information. 2-stage designs: high-recall cluster-models & IR retrieval (Candidate Generation), then high-precision matching-models. Text-categorization modeling using Transformer-based embeddings, static-embeddings, and traditional, TFIDF approaches. Feature-engineering on text. Text-similarity modelling on Product entities using NLP techniques and embeddings.Neural-network modeling: BERT; Other Transformer-based models (S-BERT, etc); CNN models (PyTorch, Tensorflow, HuggingFace).Traditional ML modeling: Random-forest, SVM, SGD, Multinomial Logistic Regression (SkLearn), Boosted-Trees (XGBoost).Transformer-based text-embeddings: using BERT, Transformer-based models.Static text-embeddings: using Fasttext, Word2Vec, Glove, Doc2Vec.• ML Product-matching streaming service (2 mo): ML inference w/ message-streaming arch (NATs, FastAPI, Redis, SQL, Kubernetes)• ML Product-similarity service (2 mo): REST API with Interfaces for Operations (Python FastAPI, React (TypeScript))

Masters in Computer Science and ResearchRelevant Coursework: Graduate CS, UCLA: • Machine Learning (AI) • Automated Reasoning (AI) • Natural Language Processing (AI,IR) • Bayesian Networks (AI) • Neural Networks & Deep Learning (AI) • Reinforcement Learning (AI) • Probabalistic Programming & Relational Learning (AI,PL) • Static Program Analysis (PL) • Type Theory & Prog.Lang(PL) • Big Data Systems (IR) • Linear Programming (Optimization) • Applied Probability (Stats)Upper-division CS, UCLA: • Algorithms and Complexity • Artificial Intelligence • Databases & DBMS • Datamining • Formal Languages & Automata Theory (Computability) • Prog. Languages (PL) • Compilers • Operating Systems • Computer Networks • Computer Graphics • Web ApplicationsUpper-division Statistics, UCLA: • Probability • Mathematical Statistics • Linear Models (MLR) • Data Analysis & Regression • Design & Analysis of Experiments • Statistical Models & Data Mining • Optimization for Statistics • Monte Carlo Methods (Audit)

Next 2.5 years: Backend Engineer, Core Technology team:• Design of custom, high-performance backend infrastructure (Java, Node.js)• Product job-queue service (Java, SQL)• High-performance streaming systems (Kafka, Redis, Java, Node.js, Docker, Socket.io)• Production log infrastructure (100 prod servers, TBs/day prod data) (Java, AWS)• Telemetry front-end app for core-infra (React, Flux, Node.js, JavaScript, AJAX, Material Design)• Customized SFTP server for customer data sync (Node.js, Docker, AWS)First 2 years: Full-Stack Product Engineer, working on production features and design:• Java backend design and data-model design (Java, Hibernate, Struts)• Custom high performance database controllers, query optimization, db tuning (Java, JDBC, SQL, Query Optimization)• Database entity extractor (Java, SQL)• Web-based file-system file deletion scheduler (Java, SQL)• Front-end design: product pages, word-search game (React, JavaScript, AJAX)

• Integrated circuit designer in custom digital and VLSI ASIC design for ultra low-power & low-voltage near-threshold circuits and CPU.• Software & scripting development for near-threshold circuit design and characterization.• Taped-out in 28nm.

• Course: EECS 216, Signals and Systems • Course topics: CT LTI systems, Freq. resp, filtering, Fourier and Laplace, basic analog comm and feedback control • Taught recitation (discussion) sections, 100 students.

• VLSI and custom digital integrated circuit design for on-chip variation sensing, tuning, and mitigation. • Ultra low power design in sub and near threshold.• VLSI tool design and automation.• Taped out in 180nm, 130nm, 90nm, 65nm, 45nm.