Featured Projects

Here is a selection of data science and AI projects I've delivered. Each one shows how I take a business problem from raw data to a system running in production, and the impact it had.

• Predictive Lead Scoring with >$1M Impact

  ---

  An enterprise lead-scoring model built end-to-end in PySpark on GCP. The top 30% of scored leads capture 85% of all conversions, driving more than $1M in business impact.

• RAG AI Tutor & LLM Evaluation Framework

  ---

  A production RAG service answering learner questions on university courses, plus the evaluation framework with hallucination-risk controls that secured enterprise AI contracts with Stanford, MIT and Berkeley.

• Search-Signal Data Pipeline: $350K in Conversion Value

  ---

  An end-to-end pipeline on SQL, BigQuery and Google APIs that turns high-intent search signals into a reusable feature and modeling layer, generating $350,000 in organic conversion value.

• Unsupervised NLP Topic Modeling for Classical Arabic

  ---

  An independent multi-phase NLP pipeline over a 6,236-verse classical Arabic corpus using transformer embeddings, UMAP, HDBSCAN and topic modeling, with automated PDF reporting.