Kasey Naqvi
I’m a senior Machine Learning Engineer with 6+ years of software engineering experience building and deploying production-grade ML systems that serve millions of users.
About
I’m a senior Machine Learning Engineer with 6+ years of software engineering experience building and deploying production-grade ML systems that serve millions of users. I specialize in ML Ops, deep learning optimization, and generative AI, and I’ve delivered NLP, computer vision, recommendation, and fraud detection solutions across fintech and healthcare.
I enjoy taking models from research to real-world reliability—cutting inference latency, improving uptime, standardizing feature pipelines, and setting up monitoring and A/B testing so teams can iterate safely. I work fluently across Python and modern deep learning stacks (PyTorch, TensorFlow) and cloud-native ML platforms on AWS and GCP, building pipelines that reduce time-to-deploy and keep models performing as data changes.
- 7 years' experience
- Joined October 2026
Industry experience
- Financial Services
- Healthcare
- Software & Internet
Work experience
Architected and deployed an NLP-based document processing pipeline using BERT on AWS SageMaker, achieving 94% accuracy and reducing manual review time and operational costs. Led an MLOps initiative implementing CI/CD with GitHub Actions, Docker, and Kubernetes, reducing deployment time from two weeks to four hours and reaching 99.9% uptime. Optimized deep learning models for real-time inference using TensorRT and ONNX quantization, cutting latency by 65% while maintaining accuracy for fraud detection on daily transactions. Built a feature engineering platform using Apache Spark and Feast to standardize features across five teams, improving reproducibility and experiment velocity by 3x. Mentored a team of four engineers on monitoring and A/B testing, and established a model governance framework to support financial regulatory compliance.
Partnered with Logic League to develop a computer vision system for medical image analysis using PyTorch and ResNet architectures, achieving diagnostic accuracy for early-stage disease detection and deploying to multiple healthcare facilities. Designed a real-time recommendation engine using collaborative filtering and deep learning, increasing user engagement through daily personalized recommendations. Implemented MLOps infrastructure on GCP Vertex AI with automated retraining pipelines, reducing model drift incidents by 90%. Engineered data pipelines processing 10TB+ daily healthcare data using Airflow, BigQuery, and Dataflow with HIPAA-compliant handling and sub-second query performance. Conducted A/B tests and statistical analyses on 15+ ML models, translating results into product decisions that improved patient outcome predictions.
Built a predictive analytics platform using scikit-learn and XGBoost for credit risk assessment, processing loan applications 10x faster and reducing default rates by 15%. Developed ETL pipelines in Python, SQL, and Apache Spark to consolidate data from 12+ sources into a unified warehouse powering real-time business intelligence dashboards. Engineered fraud detection models using XGBoost and deep learning to process transactions and prevent losses. Containerized legacy applications with Docker and orchestrated with Kubernetes, improving deployment reliability and reducing infrastructure costs by nearly 40%. Collaborated with product and engineering teams to deliver eight major features on time in agile sprints by translating business needs into technical specifications.
Education
Certifications and awards
Hire Kasey Naqvi today
Post a job and invite Kasey, or contact them directly. No hidden commission.