Open to opportunities

Dinesh User

@dineshuser5

AI/ML Engineer specializing in scalable fraud, recommendation, and NLP/LLM systems with strong MLOps and low-latency deployment.

United States

Message

What I'm looking for

I’m looking to build and ship production AI/ML systems—fraud, recommendations, and LLM/RAG—where I can own MLOps, scalable data pipelines, and low-latency deployments that drive measurable business outcomes.

I’m an AI/ML Engineer with 5+ years of experience building scalable fraud detection, recommendation, and NLP solutions on large-scale datasets. I focus on production impact—turning strong modeling into reliable, low-latency systems with measurable business lift.

At PayPal, I developed fraud detection models on 500M+ transactions, improving fraud detection precision by 18% and reducing payment chargeback losses. I also built deep learning risk scoring with TensorFlow that reduced false positives by 12% while keeping sub-50ms inference latency, and I helped containerize and orchestrate deployments on Amazon EKS to achieve 99.99% production uptime.

Before PayPal, I supported recommendation and semantic search improvements at Accenture and built churn prediction work at Infosys. I bring an engineering-minded approach to MLOps and deployment using AWS, REST APIs, Docker/Kubernetes, ML lifecycle automation, and monitoring (drift detection with Evidently AI) to keep models healthy in the real world.

Experience

Work history, roles, and key accomplishments

Current

AI/ML Engineer

Current

PayPal

Aug 2024 - Present (1 year 10 months)

Developed fraud detection models on 500M+ transactions, improving fraud precision by 18% and reducing payment chargeback losses. Built PySpark/Databricks feature pipelines and real-time Kafka-driven scoring with sub-50ms TensorFlow inference, deploying on Amazon EKS and automating MLOps in SageMaker with 25% shorter ML release cycles.

Python XGBoost TensorFlow Pyspark Apache Kafka Amazon EKS AWS Sagemaker Docker Retrieval Augmented Generation (RAG)

ML Engineer

Accenture

Aug 2021 - Jun 2023 (1 year 10 months)

Contributed to recommendation models using collaborative filtering and ranking, and enhanced semantic search relevance with BERT embeddings via Hugging Face Transformers. Built training and feature workflows with Python/SQL/PySpark on AWS, supported REST/Docker deployments, and assisted with offline evaluation and A/B testing using precision@K and NDCG.

Collaborative Filtering Learning To Rank BERT Pyspark SQL scikit learn XGBoost LightGBM A B Testing

Data Scientist

Infosys

May 2020 - Jul 2021 (1 year 2 months)

Built a customer churn prediction model using Python and Scikit-learn by preparing cleaned datasets with Pandas and NumPy and performing EDA in Python and SQL. Evaluated Logistic Regression and Decision Trees using accuracy/AUC, and created Power BI reports plus basic NLP sentiment analysis from customer feedback.