Graduate-level machine learning, graph analytics, PySpark pipelines, database systems labs, and real healthcare data tooling — organized as a deeper technical layer beneath the project archive.
A mix of portfolio-ready coursework, sanitized repo artifacts, and practical data tooling from real product environments.
Linear regression workflow on 2019 NYC listings: outlier filtering, categorical encoding, geographic signal, and low/mid/high price classification.
Network science notebooks using PySpark, NetworkX, modularity, similarity measures, and social graph community structure.
The GitHub starter pack becomes the proof layer: sanitized notebooks, artifacts, requirements, and reproducible project framing.
Readable implementations of core algorithms from graduate coursework without hiding behind library magic.
Distributed analytics practice: healthcare-style EDA and student dropout classification using Spark workflows.
Redacted lab artifacts covering schema querying, execution-plan inspection, transactions, locks, and spool workflows.
Graduate notebooks focused on the decisions that happen before modeling: missingness, outliers, encoding, scaling, and transforms.
Python utility work for pharmacy record sorting and normalization during the KARE Rx Partners tenure.
Data acquisition, pre/post-processing, integrity validation, and exploratory analysis inside a healthcare AI ecosystem.
Most analytics work is not model selection. It is framing the question, cleaning the data, understanding failure modes, validating assumptions, and turning the result into something another person can trust.
Define the decision, the stakeholder, and what “better” means.
Inspect missingness, skew, leakage, data types, outliers, and distribution shape.
Encode, scale, filter, normalize, join, and document the tradeoffs.
Build interpretable baselines before chasing complexity.
Translate metrics into decisions, risks, and next actions.
Ask me about the notebooks, the graph analytics work, the database systems labs, or how I turned coursework into clean repo-ready artifacts.
