Looking for a job

Michael Hyle

@michaelhyle

Data Engineer at American Express building fraud-model streaming and RAG data pipelines.

At American Express, I own streaming ingestion and training-data delivery for fraud and risk models. I built a Python RAG grounding-data pipeline into PGVector, reducing embedding compute by 60% through incremental re-embedding.

I also led the LLM-evaluation data layer, curating analyst-reviewed datasets and tracking evaluator runs in MLflow. My work includes online/offline feature-consistency monitoring and quality gates for regulated data.

At Mepco Finance Corp, I delivered an AWS and Snowflake ACH analytics platform and automated payment reconciliation, reducing manual effort by 70%. Earlier, at Maven Wave Partners, I helped modernize a healthcare client’s warehouse on BigQuery, and at KCI Technologies, Inc. I built data pipelines for infrastructure analytics.

Experience

Work history, roles, and key accomplishments

Current

Data Engineer

Current

American Express

Apr 2022 - Present (4 years 6 months)

Powered sub-second fraud and dispute-risk decisioning on the Risk & AI Data Platform, streaming tens of millions of daily events through Kafka and Spark Structured Streaming into a Delta Lake medallion lakehouse served via CouchDB. Owned training-data delivery for fraud and risk models, designing point-in-time temporal snapshot joins in PySpark that eliminated label leakage from delayed chargeback

Kafka Apache Spark PySpark Spark Structured Streaming Delta Lake Medallion Lakehouse Architecture CouchDB MLFlow RAG Pipeline Pgvector Python Airflow Great Expectations DBT Data Lineage PII Masking RBAC PCI Like Controls

Data Engineer

Mepco Finance Corp

Jul 2020 - Apr 2022 (1 year 9 months)

Delivered the ACH Payment Analytics Platform, the company's first cloud analytics platform on AWS and Snowflake, ingesting recurring ACH origination, return, and settlement files via S3, AWS Glue, and Snowpipe. Automated daily three-way reconciliation across payment processors, banks, and the contract-servicing system, reducing manual reconciliation effort by 70%.

AWS S3 AWS Glue Snowflake Snowpipe DBT Tableau Python Terraform Airflow CI CD Data Masking RBAC

Consultant

Maven Wave Partners

Sep 2018 - Jul 2020 (1 year 10 months)

Reduced client reporting latency from next-day batch to near real time on an Enterprise Data Warehouse Modernization engagement, migrating a healthcare client's legacy warehouse to BigQuery with Dataflow (Apache Beam), Pub/Sub streaming, and Cloud Composer (Airflow).

BigQuery Dataflow Apache Beam Pub Sub Cloud Composer Airflow

Engineer

KCI Technologies, Inc.

Aug 2011 - Jul 2016 (4 years 11 months)

Built SQL and Excel/VBA-based data pipelines for an Infrastructure Asset Data & Analytics program, consolidating field inspection, asset, and project-cost data across large infrastructure programs and automating engineering analysis models and recurring client reporting.

SQL Excel VBA