Louis Wong

Data Scientist & Applied AI Engineer building agentic workflows, ML systems, and analytics platforms that move teams faster.

I combine production data work with modern AI tooling: forecasting, ETL, lakehouse architecture, Python automation, LLM tool calling, and ML systems across AWS and Azure.

Download resume LinkedIn GitHub

[email protected]

Current roleData Scientist & Applied AI Engineer

Core laneAI agents, ML systems, and analytics automation

Graduate studyM.S. Computer Science, AI & ML at Georgia Tech

LocationLos Angeles, California

Current roleData Scientist & Applied AI Engineer

Core laneAI agents, ML systems, and analytics automation

Graduate studyM.S. Computer Science, AI & ML at Georgia Tech

LocationLos Angeles, California

Overview

A builder who can bridge model work, data platforms, and business outcomes.

My strongest work sits where analytics engineering, machine learning, and AI-assisted automation meet. I like turning vague operational pain into systems that teams can trust and reuse.

Agentic workflows

LLM tool calling, RAG patterns, Claude Code, OpenAI Codex, MCP, and process automation.

ML systems

Forecasting, investment support, profit analysis, inventory optimization, and classification pipelines.

Data platforms

SQL, PySpark, Airflow, Azure Data Factory, Delta Lake, SQL Server, and PostgreSQL.

Executive analytics

Tableau and Power BI reporting backed by clean data models and automated KPI tracking.

Agentic system architecture

A modular view of agentic AI in business systems.

The pattern is simple to reason about: business events trigger an AI agentic system, the system plans with a model and context, tools execute bounded work, and outputs route into review, insight, or operations.

Reference workflow

Business trigger to governed action

Useful agentic systems are not one giant prompt. They are composed from triggers, context, model reasoning, tools, routing logic, and human review.

Trigger

Business event or data signal

Scheduled jobs, webhooks, document drops, campaign changes, and operational events start the workflow.

Control

Human-in-the-loop review

Sensitive or high-impact steps can stop for approval, edits, or exception handling before the system acts.

Orchestration layer

AI agentic system

Plans tasks, calls tools, remembers context, and explains the next action.

Business-enhancing AI system

LLMs model

Database

Tools

LLMs model

GPT

Claude

DeepSeek

Context

SQL databases, business data, policy, history, and customer records.

Tool APIs

Real-time search, CRM actions, reporting endpoints, and workflow systems.

Logic

Complex routing

Rules, thresholds, model confidence, and policy checks decide whether to act, ask, or escalate.

Outcome

Business action

Actions, approvals, next-best recommendations, and the supporting insight reach the team that needs them.

Fallback

Operations alert

Alerts, audit logs, and monitoring close the loop when the workflow fails or needs attention.

Composable

Each module can be swapped as models, APIs, and business rules evolve.

Observable

Every step should produce traces, decisions, and reviewable outputs.

Governed

Controls, policy checks, and human review protect high-stakes work.

Experience

Applied AI, agentic automation, and production data systems.

I build AI, ML, and agent-driven workflows that turn business data into forecasting, decision support, process automation, and measurable operational value. My work spans media, marketing, and renewable energy, with a focus on building AI systems that reduce manual work, improve decisions, and create durable business advantage.

2022 to Present

Data Scientist & Applied AI Engineer

Paramount Global

Los Angeles, CA

Applied Claude Code, OpenAI Codex, LLM tool calling, and AI agents to move analytics work from manual handoffs into faster, more repeatable development and documentation flows.
Built ML and AI automation systems for forecasting, profit analysis, inventory optimization, and investment decision support, using Python, PyTorch, TensorFlow, AWS, and Azure to connect model output with business review.
Built ELT and ETL pipelines with Azure Data Factory and Airflow, consolidating enterprise data into lakehouse and reporting layers that support Tableau, Power BI, and executive KPI workflows.
Administered SQL Server and PostgreSQL environments across schema design, query optimization, reporting logic, and data architecture, balancing reliable operations with evolving analytics needs.

2021 to 2022

Marketing Science Analyst

eFunnel Experts

Remote

Developed A/B testing frameworks, segmentation models, and clustering workflows that helped marketing teams understand funnel behavior and improve return on ad spend.
Built forecasting and marketing mix modeling workflows for trend, seasonality, and channel analysis, turning campaign data into clearer budget and optimization decisions.

2017 to 2021

Business Analyst

FreeVolt Energy

Los Angeles, CA

Conducted time series analysis to forecast renewable energy production, giving stakeholders a clearer planning view for resource allocation and project timing.
Analyzed financial loan data to identify stronger target customer profiles, supporting better application strategy and more informed stakeholder conversations.

Highlighted projects

Applied AI projects built around real workflows.

These projects show how I think: connect models to tools, ship a useful interface, and keep the output clear enough for people to review.

Terminal AI Agent

Built a terminal-based AI agent with ElevenLabs audio, GPT models, callback functions, headless search automation, web scraping, and long-context summarization. This was built before agentic LLMs became widely popular, validating the direction modern AI development has since moved toward.

Early agentic designVoice input/outputHeadless search

Pioneer agentic prototype

AI Document Extraction Agent

Built a PDF-to-Excel agent that detects tables, structures extracted data, previews results, and generates review-ready XLSX workbooks for invoices, statements, reports, and pricing sheets.

Table detectionReviewable previewsXLSX generation

View project

AWS End-to-End ML Pipeline

Built a SageMaker healthcare classification pipeline covering raw data ingestion, training, deployment, and stroke risk prediction with scalable cloud-native tools.

SageMakerModel deploymentHealthcare classification

View project

Research and education

Academic AI foundation, applied in industry.

My background pairs graduate AI and machine learning study with research experience in multimodal deep learning.

Multimodal Deep Regression on Video Content

Research lead for a multimodal architecture combining video encoding, audio processing, and natural language representations to predict TikTok content success.

View PRIP 2023 proceedings

Education

M.S. in Computer Science

AI & Machine Learning, Georgia Institute of Technology

B.S. in Computer Science

Western Governors University

Certifications

AWS Certified Machine Learning Specialty Microsoft Certified Azure AI Engineer Associate

Anthropic certified: Claude Code in Action

Data Science Professional by IBM

Machine Learning Professional by IBM

Marketing Analytics Professional by Meta

CompTIA Project+

ITIL 4 Foundation by AXELOS

Resume

One-page resume, ready for interview loops.

Download the PDF for applicant tracking systems, recruiter forwards, and hiring manager review.

Download PDF Email Louis

LLMsAI agentsAgentic workflowsRAGTool callingMCPOpenAI APIClaude CodeOpenAI CodexPythonPyTorchTensorFlowSQLPySpark

Open PDF

Let's talk

Ready to upgrade your business with agentic AI and integrated data platforms?

I help turn scattered data, manual workflows, and AI ideas into practical systems that automate decisions, surface insights, and keep teams moving with confidence.

[email protected]Connect on LinkedIn