Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I'm Mounesh Rayalla, a Machine Learning Engineer based in Missouri with 3+ years building production ML pipelines and feature platforms for Fortune 500 clients across GCP, AWS, and Azure. I've progressed from data engineering foundations into owning end-to-end ML workflows, delivering measurable business impact through churn modeling, predictive maintenance, and real-time streaming systems. I focus on closing the gap between raw data and reliable model output in fast-moving consulting environments where data quality and feature consistency directly determine production success. I thrive on turning messy data into trustworthy features and scalable pipelines that power data-driven decisions.…Hi, I'm Mounesh Rayalla, a Machine Learning Engineer based in Missouri with 3+ years building production ML pipelines and feature platforms for Fortune 500 clients across GCP, AWS, and Azure. I've progressed from data engineering foundations into owning end-to-end ML workflows, delivering measurable business impact through churn modeling, predictive maintenance, and real-time streaming systems. I focus on closing the gap between raw data and reliable model output in fast-moving consulting environments where data quality and feature consistency directly determine production success. I thrive on turning messy data into trustworthy features and scalable pipelines that power data-driven decisions.

Mounesh Rayalla

Data Scientist, AI Engineer, Full Stack Developer, +3





Hi, I'm Mounesh Rayalla, a Machine Learning Engineer based in Missouri with 3+ years building production ML pipelines and feature platforms for Fortune 500 clients across GCP, AWS, and Azure. I've progressed from data engineering foundations into owning end-to-end ML workflows, delivering measurable business impact through churn modeling, predictive maintenance, and real-time streaming systems. I focus on closing the gap between raw data and reliable model output in fast-moving consulting environments where data quality and feature consistency directly determine production success. I thrive on turning messy data into trustworthy features and scalable pipelines that power data-driven decisions.…Hi, I'm Mounesh Rayalla, a Machine Learning Engineer based in Missouri with 3+ years building production ML pipelines and feature platforms for Fortune 500 clients across GCP, AWS, and Azure. I've progressed from data engineering foundations into owning end-to-end ML workflows, delivering measurable business impact through churn modeling, predictive maintenance, and real-time streaming systems. I focus on closing the gap between raw data and reliable model output in fast-moving consulting environments where data quality and feature consistency directly determine production success. I thrive on turning messy data into trustworthy features and scalable pipelines that power data-driven decisions.

Available to hire

Hi, I’m Mounesh Rayalla, a Machine Learning Engineer based in Missouri with 3+ years building production ML pipelines and feature platforms for Fortune 500 clients across GCP, AWS, and Azure. I’ve progressed from data engineering foundations into owning end-to-end ML workflows, delivering measurable business impact through churn modeling, predictive maintenance, and real-time streaming systems.

I focus on closing the gap between raw data and reliable model output in fast-moving consulting environments where data quality and feature consistency directly determine production success. I thrive on turning messy data into trustworthy features and scalable pipelines that power data-driven decisions.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

ML Engineer at Accenture

January 1, 2025 - Present

Led end-to-end ML feature engineering and pipeline development: built 40+ reusable features stored in Databricks Feature Store with PySpark, enabling consistent model training across 3 client units and reducing feature engineering overhead by 60%. Designed and executed champion-challenger churn model A/B tests using MLflow with a 70/30 traffic split; the challenger improved precision by 6% and was rolled out. Performed business metric analyses by linking model outputs to Synapse revenue data; identified top-decile churners representing 34% of at-risk contract value, guiding retention priorities. Built Azure Data Factory pipelines ingesting from 11 source systems into a Delta Lake on ADLS Gen2, supporting a Synapse-based reporting layer for 300+ users. Optimized Delta Live Tables bronze-to-silver transformations with Z-order clustering, cutting training time on a 22M-record daily dataset from 4.2h to 2.6h. Implemented Great Expectations data quality checks across 11 feeds, flagging 3,00

Data Engineer at Fractal Analytics

January 1, 2022 - December 1, 2023

Built a 2-week maintenance demand forecasting model using Facebook Prophet on GCP Dataproc, chosen for its ability to handle irregular IoT seasonality without manual stationarity tuning, achieving 12.4% MAPE and reducing reactive repair incidents by an estimated 22%. Improved equipment failure F1 from 0.72 to 0.83 across 6 MLflow-tracked experiments by adding rolling window and lag features to a Random Forest classifier. Replaced a 4-hour batch process with a real-time streaming pipeline on Apache Kafka and GCP Pub/Sub to feed near real-time sensor events into the predictive maintenance model, achieving sub-2-second latency. Caught drift in 4 of 12 sensor channels before they reached the ML pipeline by running KS and chi-square tests across 3 upstream feed migrations, preventing two silent model degradation incidents and triggering schema validation updates in Great Expectations. Built PySpark feature pipelines on GCP Dataproc covering 8M+ records, implementing lag features and tempora

Jr. Data Engineer at Fractal Analytics

January 1, 2022 - December 1, 2023

Designed time-series forecasting models using Prophet on GCP Dataproc to predict maintenance demand with 18 months IoT history; achieved 12.4% MAPE on holdout and informed scheduling to reduce reactive repairs by ~22%. Built real-time streaming pipeline on Apache Kafka and GCP Pub/Sub with sub-2-second latency; led drift-detection experiments (KS/Chi-square) across sensor feeds and updated schema validations to prevent degradation. Created ML-ready feature pipelines with PySpark and Spark MLlib and consolidated 9 source systems into a GCP BigQuery data warehouse for self-serve feature extraction.