Open to opportunities

Andrew Han

@andrewhan

Message

Staff ML engineer specializing in production MLOps, real-time data pipelines, and agentic LLM systems at scale.

United States

Message

What I'm looking for

I’m looking for a team where I can build production-grade ML and modern AI (LLMs, RAG, agentic systems) with real-time data pipelines, strong MLOps, and measurable impact—using observability, drift detection, and experimentation to ship reliable systems.

I’m a Staff Machine Learning Engineer with 10+ years of experience in Data Science, Data Engineering, and MLOps, building large-scale ML systems across fintech, social media, and enterprise platforms. I focus on designing real-time data pipelines, ML infrastructure, and production-grade model systems that perform reliably at scale.

My hands-on expertise spans Spark, Kafka, AWS/GCP, and end-to-end MLOps with MLflow and Kubernetes. Recently, I’ve been focused on modern AI systems—LLMs, RAG pipelines, and agentic AI architectures—so enterprises can automate workflows with stronger grounding and safer operations.

I’ve delivered agentic and data-driven outcomes end-to-end, from healthcare AI agent systems to large-scale personalization and fraud detection. In production, I prioritize observability, drift detection, monitoring, and experiment-driven iteration (including A/B testing), so models stay accurate as data and environments change.

Experience

Work history, roles, and key accomplishments

Staff Machine Learning Engineer

Salesforce

Jan 2026 - Jun 2026 (5 months)

Architected and developed an AI agent system for healthcare workflows on Salesforce Health Cloud/Data Cloud, automating clinical and administrative operations. Built healthcare data pipelines and RAG systems, and implemented production MLOps for agent deployments and monitoring.

Python REST APIs BigQuery Data Pipelines Streaming And Batch ETL Patient Master Record fastAPI RAG Embeddings MLFlow Kubernetes Prometheus Docker Audit Logging

Senior ML Engineer

Alchemy

Nov 2022 - Dec 2024 (2 years 1 month)

Developed production graph-based machine learning systems for blockchain fraud analytics, including wallet clustering, entity resolution, and transaction risk scoring. Implemented real-time streaming feature pipelines, end-to-end MLOps with continuous retraining, and monitoring/drift detection for non-stationary network behavior.

Monitoring Drift Graph ML Wallet Clustering Entity Resolution Structural Graph Features Representation Learning Event Driven Architecture MLOps Automated Data Ingestion Feature Generation Model Training Deployment Anomaly Detection Statistical Analysis

Senior ML Engineer

Cash App

May 2021 - Oct 2022 (1 year 5 months)

Built real-time fraud detection and risk scoring models for high-volume transaction streams using PyTorch and LightGBM, with version tracking in a model registry. Implemented graph-based anomaly detection, low-latency inference, production monitoring for drift/performance degradation, and A/B testing to improve accuracy while reducing false positives.

Monitoring Drift PyTorch LightGBM Model Versioning Feature Importance Analysis Fraud Prevention A B Testing Fraud Threshold Optimization Anomaly Detection

Product Data Scientist (Classification)

Square

Aug 2017 - Apr 2021 (3 years 8 months)

Led a large-scale product categorization initiative that classified 50M+ products into 10K+ taxonomy categories, reducing manual tagging by 90%. Built supervised text classifiers and image embedding pipelines, and developed scalable Spark/Hadoop data processing to improve dataset consistency and batch processing time.

SQL Pandas TF IDF Logistic Regression SVM Random Forest XGBoost TensorFlow ResNet Apache Spark Hadoop Supervised Learning N Gram

Data Scientist (Marketplace)

Uber

Jan 2016 - Jul 2017 (1 year 6 months)

Developed predictive models for ETA, driver acceptance probability, and cancellation risk using TensorFlow and probabilistic modeling. Built distributed data pipelines for spatiotemporal signals, executed large-scale A/B tests, and delivered KPI dashboards for marketplace performance tracking.