Adrian Lam Email & Phone Number

South San Francisco, California, US

San Francisco, California, US

Focused on data-centric strategies to maximize LLMs for use in enterprise applications including finetuning LLMs for domain specific tasks, retrieval augmented generation and text2sql

San Francisco, California, US

Built a high-quality ML pipeline to extract data from any document type using state of the art OCR + information extraction deep learning language models

Nashville, Tennessee, US

• Responsible for leading a team of data scientists to build end-to-end products from conception to deployment that have direct impact on healthcare workflows. Create and manage product roadmap for features and.
• Productionized data science pipeline to generate audit predictions from medical record PDFs. Processes 100M pages annually to reduce fraud, waste and abuse in healthcare.
• Work with business units to collect requirements, evaluate feasibility of proposals and determine financial impact of models
• Plan weekly data science and engineering stories, manage progress to remove any blockers and to ensure timely delivery of features
• Pioneered development of code repositories used across the team for large scale processing of text, deep learning pipelines, and cloud utilities
• Heavily involved with standardizing DevOps practices across the AI team for code quality, CI/CD, testing, and deployment of models and pipelines Medical Record Audit Automation:

Nashville, Tennessee, US

• Responsibilities include:Digital Assistant
• Successful Beta and Beta+ delivery of chatbot application with 20+ customer use cases to simplify access to CRM/HR information
• Developed complete NLP intent recognition and parameter extraction chatbot system using NLP4J by novel dependency tree comparison algorithm
• Integrated scoring system to rate intent recognition by passing engineered features into LibSVM
• Driving force behind innovation of multiple secondary features by guided interns to successfully create and integrate TF-IDF recommendations, Naïve-Bayes autocorrect, and fuzzy matched entity suggestionsIntelevator as.
• Worked with international team to build multiple end-to-end scenarios to support high volume IoT data by combining data generation, Flume, Kafka, Spark Streaming, SAP HANA Vora, and SAP HANA

Research and development for strategic initiatives at SAP

Atlanta, Georgia, US

• PhD Dissertation: Development of a Combined Angiography and Late Gadolinium Enhancement (LGE) MR Sequence
• Used Siemens IDEA/ICE platform to create a cardiovascular pulse sequence to acquire automatically co-registered angiography and LGE scar images
• Designed and implemented algorithms for semi-automated myocardial segmentation and image co-registration
• Software pipeline developed in MATLAB to reconstruct MRI images into clinically relevant 3D models and 2D projections
• Produced coronary vein and myocardial scar maps to allow pre-procedural planning of pacemaker lead implantation to optimize patient benefit
• Pioneered development of pulmonary vein scar maps to quantify and compare extent of electrical isolation by cryo-balloon/RF ablation

Phoenix, AZ, US

• Intensive, 7-week data science course with a 2% acceptance rate that equips scientists and engineers with advanced degrees the tools to extract predictive capabilities from big data. Projects developed include:
• Extracted link statistics on a subset of Wikipedia using Amazon Web Service's (AWS) Elastic Map Reduce (EMR)
• Models developed to predict the release date of a generic counterpart of a brand name drug to impact monthly sales forecasting
• Analyzed user and post statistics on Stack Overflow data with Apache Spark
• Predicted Yelp ratings with machine learning algorithms and performed sentiment anaylsis on reviews
• Developed a graphical network of New York's social elite using networkx

• Learn about fluorescent lifetime imaging and spectroscopy systems. • Investigate cellular and DNA damage caused by laser-induced fluorescence

Phd, Bioengineering

Bs, Biomedical Engineering