Looking for a job

Malik Hamza

@malikhamza3

Senior AI/ML Engineer at MarLabs Inc who shipped a clinical RAG platform that cut document retrieval from 8 minutes to under 20 seconds.

United States

Message

At MarLabs Inc, I architected and shipped a production RAG platform over clinical policy, compliance, and training documents, cutting retrieval time from about eight minutes to under 20 seconds. I also designed a multiagent system for provider credentialing and license verification that reduced manual review effort by about 65%.

At Caylent, I delivered production machine learning solutions, including an eCommerce recommendation engine that increased click-through rate by 24% and average order value by 11%. Earlier, as a Data Scientist at Weston Chase, I built credit risk models that improved the Gini coefficient by 22% over the incumbent scorecard.

Experience

Work history, roles, and key accomplishments

Current

Senior AI/ML Engineer

Current

MarLabs Inc

Mar 2022 - Present (4 years 7 months)

Architected and shipped a production RAG platform over 2M+ clinical documents using AWS Bedrock and Azure OpenAI, cutting retrieval time from 8 minutes to under 20 seconds. Designed multiagent Agentic AI systems and fine-tuned LLMs, reducing manual review effort by 65% and inference costs by 58%.

Machine Learning Engineer

Caylent

Jun 2020 - Feb 2022 (1 year 8 months)

Delivered 15+ production ML and data solutions for enterprise clients, including a real-time recommendation engine, demand forecasting, fraud detection, and computer vision systems. Automated infrastructure with Terraform and reduced client ML infrastructure spend by 35%.

Sagemaker DeepAR Prophet LightGBM XGBoost Isolation Forest Feast Kafka Kinesis YOLO V5 TensorRT NVIDIA Jetson BERT Terraform CloudFormation Amazon EKS Step Functions Glue Airflow A B Testing

Data Scientist

Weston Chase

Jan 2019 - May 2020 (1 year 4 months)

Built credit risk and default prediction models, anomaly detection for AML, and time series forecasting. Automated reporting and ETL pipelines, saving 25 analyst hours per week, and designed A/B tests with causal inference.