FlowCV Logo
Professional Summary
Solution Architect and AI/Data Engineering leader with 18 years of experience designing and deploying enterprise-scale data, ML, and GenAI solutions. Deep expertise in large language models (LLMs), distributed data engineering, and cloud-native architectures for healthcare, media, and AdTech. Proven record of delivering measurable business impact and regulatory-compliant solutions for Fortune 500 clients and startups. Specialized in Python, Apache Spark, Databricks, AWS, Azure, and healthcare data interoperability.
Professional Experience
Cofounder & AI Engineer, Arcs Health
  • •Led the creation of Scriber, a real-time AI system for transcribing clinical conversations using OpenAI Whisper and Llama 3.2 70B, reducing clinician note-taking time by 60% and ensuring HIPAA compliance.
  • •Built scalable, multi-format healthcare data pipelines (CCDA, HL7 v2/v3, EDI) with Azure Data Factory, Databricks, and Fabric, standardizing to FHIR for analytics and interoperability.
    • 01/2023 – PresentNew York, NY
  • •Designed and implemented a RAG-enabled patient engagement chatbot (Twilio, Llama 3.2 70B, LangChain Agentic), automating appointment scheduling and triage for 5,000+ patients/month.
  • •Developed digitized intake and document upload workflows, streamlining patient registration and insurance verification.
  • •Architected secure, regulatory-compliant data workflows, integrating Azure AD authentication and API Management for privacy and access control.
    • Sr. Data Engineer, NBC News
  • •Automated CCPA/GDPR Data Subject Request (DSR) processing using Airflow and Databricks Delta Lake, reducing manual workload by 85% and ensuring regulatory compliance.
  • •Optimized Spark streaming jobs for MSNBC/CNBC, resolving data skew and DAG lineage issues, improving real-time processing efficiency by 35%.
    • 09/2024 – 01/2025Englewood Cliffs, NJ
  • •Developed real-time BI dashboards (Looker, AWS Kinesis, Kafka, EMR) for subscription and viewership analytics, reducing latency to 5 minutes.
  • •Led Terraform-based CI/CD deployments and orchestrated complex data workflows using Airflow.
    • Staff Software Engineer, Preveta
  • •Architected HIPAA-compliant data pipelines for major EHR/EMR integrations using Python, Databricks, and Azure Synapse.
  • •Designed secure ingestion platforms for bi-directional data flows, supporting PowerBI analytics and regulatory reporting.
    • 03/2023 – 02/2024Los Angeles, CA
  • •Developed robust Python libraries for HL7, C-CDA, and EDI standardization, enabling rapid onboarding of healthcare partners.
    • Senior Solution Architect, The Trade Desk
  • •Designed and implemented custom ML models and CI/CD pipelines (Python, Spark) for ad campaign optimization, delivering up to 35% ROI improvement for clients such as McDonald's and Bayer.
  • •Led multi-million-dollar data science projects, including custom bidding algorithms and supply path optimization, reducing cost per viewable impression by 12%.
    • 11/2019 – 10/2022New York, NY
  • •Built real-time analytics solutions integrating DSP, ad server, and audience data for programmatic marketing automation.
    • Founding Director, Intellinum Analytics Inc
  • •Led AI/data engineering projects for AdTech, healthcare, and retail clients, from ideation to delivery.
  • •Optimized Spark jobs on Kubernetes, reducing operational costs by 70–80% and accelerating development cycles.
    • 02/2017 – 01/2023New York, NY
  • •Designed and deployed ensemble ML models for campaign optimization, achieving up to 112% test campaign performance uplift.
    • Research Staff, IBM Research
  • •Developed city analytics and customer segmentation solutions using Python ETL, Spark, and IBM Big Insights.
  • •Built predictive models for healthcare and media, delivering actionable insights from large-scale, multi-source data.
    • 2015 – 2018Yorktown Heights, NY
  • •Led real-time analytics platform development for telecommunications, supporting campaigns for 35M+ users.
    • Solutions Architect, EMC
  • •Provided backend engineering and infrastructure consulting for clients including Monsanto, specializing in VCE converged infrastructure and Python-based automation.
    • 2014 – 2015New York, NY
    Education
    M.S. in Computer Science, New York University
    2012 – 2014NY
    B.S. in Computer Science, Donghua University
    2005 – 2009China
    Technical Skills
    Languages

    Python

    Scala

    Java

    SQL

    JavaScript

    Go

    Rust

    R

    C#

    Cloud

    AWS (EMR, Lambda, S3, SageMaker, EC2)

    Azure (Data Factory, Synapse, OpenAI, Kubernetes, APIM)

    GCP (BigQuery, Dataflow, Vertex AI)

    Healthcare Data

    FHIR

    HL7

    C-CDA

    EDI

    Security

    Azure AD

    Azure API Management

    OAuth2

    Frameworks

    PyTorch

    Spark

    Scikit-learn

    Transformers

    TensorFlow

    Semantic Kernel

    FastAPI

    Data

    Databricks

    Delta Lake

    Hive

    Redshift

    Snowflake

    Kafka

    Airflow

    dbt

    DevOps

    Docker

    Kubernetes

    Terraform

    Linux

    CI/CD

    Recent Projects
    Enterprise Agentic Automations Platform, 2025
  • •Architected an enterprise-ready AI research agent by porting an open-source solution to Microsoft's Semantic Kernel, enabling modular LLM agent orchestration and plugin-based skill ecosystem.
  • •Integrated Azure AD authentication and custom API Management (APIM) for secure, role-based access and rate-limiting, supporting 10,000+ research queries per day.
  • •Implemented the MCP Server/Client using Python MCP SDK to support calling various tools.
  • •Implemented memory/context management for LLMs, optimizing research workflows and context window usage.
  • •Containerized the platform for flexible deployment across Azure Kubernetes Service (AKS) and hybrid environments.
  • •Developed specialized reporting modules for insurance, safety, and financial research, supporting custom compliance workflows.
  • •Lessons learned: MCP via SSE, LLM context management, authentication flow, and streaming architecture patterns.
    • AI Evaluation & Governance Framework, 2025
  • •Built a custom evaluation system for LLM agents using Azure AI Evaluation SDK, tailored for enterprise deployments behind Azure API Management (APIM) with custom authentication and header requirements.
  • •Developed APIM-aware evaluators for metrics including groundedness, factual accuracy, relevance, contextual precision, faithfulness, and fluency, supporting JSON-based scoring and reasoning.
  • •Implemented a manual evaluation loop for robust aggregation and reporting, overcoming SDK limitations with APIM endpoints.
  • •Integrated caching and error handling for stability and cost efficiency; enabled detailed pass/fail analytics and compliance reporting.
  • •Provided actionable insights to improve prompt engineering, retrieval strategies, and system reliability.
    • GPU-Accelerated RAG Pipelines, 2025
  • •Designed and delivered a GPU-accelerated Retrieval-Augmented Generation (RAG) system for enterprise document intelligence, achieving 5x speedup over CPU-based processing.
  • •Implemented multi-GPU orchestration for parallel document parsing, embedding generation, and chunking, scaling to 50+ pages/sec.
  • •Developed memory-aware batching and dynamic chunking strategies to optimize GPU utilization and prevent out-of-memory errors.
  • •Built robust checkpointing and verification systems to ensure data integrity and enable seamless failover to CPU as needed.
  • •Containerized the pipeline for reproducible deployment; integrated monitoring for GPU utilization and performance metrics.
  • •Realized dramatic reductions in processing time and cost per document, enabling real-time analytics on petabyte-scale document collections.
    • Publications
    ICCCS 2018

    Ding X., Yan C., Zhao Y., Yang Z. (2018). Efficient Processing of TopK Dominating Queries on Incomplete Data Using MapReduce. *ICCCS 2018*, Cloud Computing and Security, pp. 78–89.

    12/01/2017
    Certifications
    Databricks Certified Machine Learning Professional

    https://credentials.databricks.com/aa12012c-d1ae-195-a99c-2b95d99ffa2#acc.apZlwUGe

    AWS Certified AI Practitioner

    https://www.credly.com/badges/1351a19d-0020-3f3-8fa0-16d8583bceb0/public_url

    Databricks Certified Data Engineer Professional

    https://credentials.databricks.com/6abe7e2-163a-3ad-ab2f-bee8999a90f#acc.sgXrZzbq

    Microsoft Certified: Azure AI Fundamentals

    https://learn.microsoft.com/en-us/users/jonroosevelt/transcript/dlozriqzx8g9wm

    AWS Certified Machine Learning – Specialty

    https://www.credly.com/badges/33bd7b0-5301-7b-b91d-68b5275e627/public_url

    Core Competencies
    GenAI & LLMs

    RAG, Llama Index, LangChain/LangGraph/LangSmith, LLM fine-tuning, evaluation/governance, agentic frameworks, MCP, prompt engineering, vector databases, agentic memory

    Data Engineering

    Spark, Databricks, Delta Lake, Airflow, Kafka, Hive, Redshift, Snowflake, AWS EMR, Azure Data Factory, dbt, medallion architecture, structured streaming

    Cloud Platforms

    AWS (EMR, Lambda, S3, SageMaker, EC2)

    Azure (Data Factory, Synapse, OpenAI, Kubernetes, APIM)

    GCP (BigQuery, Dataflow, Vertex AI)

    Healthcare Data

    FHIR, HL7 (v2/v3), C-CDA, EDI, HIPAA/GDPR/CCPA compliance, EHR/EMR integration, data normalization, clinical NLP

    Machine Learning

    PyTorch, Scikit-learn, Transformers, TensorFlow, ML pipelines, CI/CD, model evaluation, experiment tracking, MLflow

    DevOps & Security

    Docker, Kubernetes, Terraform, Linux, CI/CD automation, Azure AD, API Management, secure API design, identity integration