Data Science
Extracting value from complex data through statistics, visualization, segmentation, and advanced analytics.
PythonSQLPandasStatisticsVisualization
Jason Lai
Data Science | Machine Learning | Applied AI
Turning Data into Models, Insights, and AI Products
I'm Jason Lai, a business-minded Data Scientist with a background in statistics, computer science, and management analytics. I build data-driven solutions across analytics, machine learning, and applied AI, transforming complex data into business insights, predictive models, and intelligent products.
Python
SQL
MLflow
What I Do
Data science, modeling, analytics, and AI products in one workflow.
I connect business context with technical execution, moving from messy data to insights, validated models, and usable deployed tools.
Machine Learning
Building predictive models with rigorous validation, feature engineering, and business-aware evaluation.
PyTorchTensorFlowLightGBMXGBoostAnomaly Detection
Business Analytics
Turning analysis into decisions through KPI design, stakeholder-ready insights, dashboards, and case framing.
KPI DesignTableauExcelSegmentationStorytelling
Applied AI
Designing practical AI products with RAG, agents, tool use, guardrails, and evaluation workflows.
RAGLangGraphFAISSFastAPIDocker
Profile notes
- Rotman Master of Management Analytics candidate
- UBC Computer Science and Statistics background
- Experience across financial services, industrial technology, and retail
Projects
Featured work, filterable by role.
Showing 4 / 10
FeaturedAI / LLM Systems
Coinbase Support Agent
Built an agentic customer support system using LangGraph, hybrid RAG, tool use, memory, and safety guardrails to resolve domain-specific support workflows.
LangGraphRAGFAISSBM25
100% task success97%+ grounded response quality60+ documents indexed
FeaturedAI / LLM Systems
AML Agentic Intelligence Engine
Designed a governed multi-agent AML investigation system that converts noisy transaction alerts into investigator-ready case intelligence for triage, risk reasoning, evidence assembly, and SAR/STR drafting.
LangGraphRAGGuardrailsAgent State
15% high-risk detection improvement20% invalid alert reductionHuman-in-the-loop review
FeaturedData Analytics
Retail Profitability & Customer Analytics
Analyzed multi-brand retail performance to identify profitability drivers, diagnose underperformance, and recommend business actions across store and customer segments.
PythonSQLTableauExcel
12% operating cost reduction9.8% mid-year sales improvement8 brands analyzed
NDA ProtectedLearn More
FeaturedMachine Learning
Art Style Classification
Built a deep learning pipeline for classifying art movements using CNN and transfer learning experiments, with model evaluation and interpretability through Grad-CAM.
TensorFlowKerasCNNViT
90%+ baseline CNN accuracyModel comparisonVisual interpretability
NDA ProtectedLearn More
Experience
Applied data science in financial services, industrial technology, and retail.
A compact timeline of roles where business impact, modeling, analytics, and AI systems came together.
January 2026 - Present
Data Scientist Intern, AML
Scotiabank
Toronto, Ontario
Business impact
Improved AML alert quality, reduced false positives, and supported investigator-ready case intelligence.
PythonPyTorchIsolation ForestAutoencoder
Technical work
Built Isolation Forest, Autoencoder, and semi-supervised models across transaction typologies; engineered 50+ behavioral and time-series features; prototyped a governed multi-agent AML investigation system with LangGraph, RAG, guardrails, audit logging, and evaluation.
PythonPyTorchIsolation ForestAutoencoderLangGraphRAGLLM-as-JudgeFeature Engineering
October 2024 - May 2025
Data Scientist Intern
Siemens Digital Industries
Beijing, China
Business impact
Built ML and AI workflows to improve intent triage, document search, and cross-department knowledge access.
TensorFlowLightGBMFAISSRAG
Technical work
Preprocessed text and manufacturing datasets, used SMOTE-Tomek for class balancing, implemented CNN + Bi-LSTM + LightGBM model, built FAISS knowledge base, Dockerized ETL, and developed RAG chatbot with localized DeepSeek 32B.
TensorFlowLightGBMFAISSRAGDockerAWSPythonDeepSeek
July 2024 - October 2024
Data Analyst Intern
Pou Sheng International
Dalian, China
Business impact
Analyzed retail store and brand performance to identify profitability drivers and recommend operational actions.
PythonSQLTableauExcel
Technical work
Built profitability and customer analytics frameworks using ARPU, repurchase rate, sales per square foot, segmentation, and visualization.
PythonSQLTableauExcelKPI AnalysisCustomer Analytics
Awards
Case competitions that connect analytics with executive storytelling.
These experiences show business framing, stakeholder communication, and data-backed recommendation work.
Data Analysis / Tech Consulting
1st Place - Calian Case Competition
Analyzed a real-world business problem, identified key operational and analytical drivers, and delivered a data-backed recommendation strategy to judges.
My role: Data analysis, insight generation, business framing, and presentation support.
CRM Data AnalysisCustomer ClusteringBusiness Problem SolvingStorytellingTeamworkExecutive Presentation
Credit Risk / Financial Analytics
2nd Place - Scotiabank Case Competition
Developed a structured credit risk analysis approach, identified meaningful risk drivers, and translated findings into business recommendations.
My role: Risk analysis, model reasoning, business recommendation, and presentation.
Financial AnalyticsCredit RiskData StorytellingModelingStakeholder Communication
Public Sector / Causal Inference
2nd Place - Toronto Police Service Case Competition
Worked on a real-world analytics case involving public-sector decision-making, analytical framing, and team-based recommendation development.
My role: Data analysis, research, insight development, and presentation.
AnalyticsCausal InferenceStrategyCommunicationTeam Collaboration
Resume
Resume
Looking for the full version of my experience, projects, and technical background? Download my latest resume below.
Contact
Let's Connect
I'm open to Data Scientist, AI/ML Engineer, Applied AI, and Analytics opportunities where I can combine data analysis, machine learning, and business problem solving.