Open to opportunities

Steve Liu

@steveliu

Message

Machine Learning Engineer with 10+ years delivering production-grade AI at scale.

United States

Message

What I'm looking for

I’m looking to build and scale production AI systems end-to-end—ML engineering + MLOps—focused on multimodal/LLM use cases, reliable CI/CD, model monitoring (drift + health), and measurable business outcomes.

I’m a Machine Learning Engineer with 10+ years delivering production-grade AI systems across enterprise, consumer, and healthcare. I design and manage large-scale AI platforms focused on reliability, scalability, and continuous model performance.

I’ve led end-to-end ML engineering pipelines in Python and PyTorch, integrating NLP and transformer architectures to process over 10 million customer records—delivering a 42% improvement in automated diagnosis accuracy while maintaining compliance on AWS. I also build multi-modal ML solutions that increase detection performance (25% increase in early behavior detection across 500,000 customer interactions).

My work emphasizes MLOps and model health: I build automated retraining and ML monitoring using MLflow Tracking, Kubeflow Pipelines, and Vertex AI pipelines, detecting data drift and maintaining accuracy above 95%. I deliver backend ML systems and RESTful APIs that scale to thousands of requests per second through robust CI/CD validation gates.

I’m passionate about transforming cutting-edge research into high-impact product outcomes—whether that’s multimodal systems, edge intelligence, or explainable, compliant AI that supports critical decision-making.

Experience

Work history, roles, and key accomplishments

Current

Senior Software Engineer

Current

Abridge

Feb 2022 - Present (4 years 3 months)

Designed and deployed end-to-end ML pipelines in Python/PyTorch for NLP and transformer architectures, processing 10M+ customer records and improving automated diagnosis accuracy by 42% while maintaining compliance via AWS CI/CD. Built multimodal predictive analytics with a 25% increase in early behavior detection, implemented retraining and model health monitoring (data drift detection) keeping a

Python PyTorch AWS CI CD NLP Transformers MLFlow Vertex AI Apache Spark

Senior AI Engineer

Iterative Health

Jun 2019 - Jan 2022 (2 years 7 months)

Designed and trained computer vision deep learning models using TensorFlow/Keras and GPU-accelerated mixed precision, improving object detection by 42% through rigorous evaluation. Built and deployed NLP workflows with Hugging Face Transformers and PyTorch Lightning, extended into multimodal models, and implemented production ML workflows with Docker, SageMaker/Vertex AI pipelines, MLflow tracking

TensorFlow Keras Pytorch Lightning Docker Sagemaker Vertex AI MLFlow Kubeflow

Software Engineer

Degreed

Sep 2016 - May 2019 (2 years 8 months)

Built and optimized large-scale search and recommendation features by blending user skills signals with content metadata, increasing relevant-result CTR by 18% and reducing null-result sessions by 42%. Rebuilt React and Java service flows to increase weekly active users by 23% and pathway conversion by 31%, implemented OAuth2.0-based API integrations, and improved reliability with tenant-safe API

Java React OAuth 2.0 REST APIs Microservices Elasticsearch SLOs Synthetic Monitoring Circuit Breakers

Junior Software Engineer

Alma

Jun 2014 - Aug 2016 (2 years 2 months)

Built healthcare onboarding, credentialing, and workflow systems that reduced time to credential by 32% and increased first-pass approvals by 18%. Developed eligibility and claims-processing workflows using ANSI X12 transaction handling with retry/reconciliation logic, improving clean-claim rate by 11 points and reducing time to payout by 2.5 days, and improved backend performance and reliability

Workflows X12 Retry Logic Caching Monitoring