End-to-end ML | Customer Segmentation | Recommender Systems | A/B Testing | GCP · Vertex AI

• Hybrid Offer Engine — lightFM + XGBoost hybrid recsys with offline & online evaluation · ranking metrics, cold-start mitigation (Coming Soon)
• A/B Testing & Causal Inference — conversion uplift pipeline with CUPED/IPSW weighting · power analysis, variance reduction (Coming Soon)
• Vertex AI Pipeline Skeleton — reproducible training & batch scoring with GCS/BigQuery · DAGs, CI/CD (Coming Soon)
• Bayesian Modeling — hierarchical model for campaign lift estimation · interpretable posteriors (Coming Soon)
• Customer Segmentation Modelling (Coming Soon)
• Customer Purchase Propensity Modelling (Coming Soon)

I work on large-scale retail analytics and loyalty-focused ML. My current interests: deep recsys, agentic evaluation, customer analytics, feature experimentation, NLP, and cost-sensitive targeting. I realize a lot of my work is stuck on company repositories which I unfortunately cannot share here. I am working on reproducing some semblance of these projects here.

• Languages: Python, SQL
• ML: XGBoost, LightGBM, RandomForest, LightFM, TensorFlow/Keras, LLMs
• Data: BigQuery, pandas, PySpark (on GCP or DataLake)
• MLOps: Vertex AI, Docker, Airflow, CI/CD
• Experimentation: A/B testing frameworks, uplift modeling

• LinkedIn: https://www.linkedin.com/in/immanueltacky
• Email: (DM me for my email)