Nathaniel N.

GenAI Platform & Foundation Architecture • Developed enterprise AI foundation architecture and platform capabilities, including Agentic AI, self-hosted LLMs, cross-LLM provider integration, and federated learning to support the full AI lifecycle. • Designed knowledge base and data architecture optimized for AI consumption. • Developed AIOps solutions with built-in AI risk and responsible AI compliance in highly regulated environment. • Collaborated with the MLOps team to enhance SDLC processes and operational frameworks, enabling scalable and reliable AIOps deployments. • Established technology stacks for AI experiments, optimization, evaluation, and monitoring. Enterprise GenAI/ML Product Development & Project Delivery • Designed and productionized enterprise-grade AI solutions including: Financial Investment Advisory AI (Agentic RAG), Audio Market Conduct Validation (Automation Workflow), HR Chatbot (Agentic) & HR Recruitment Pre-Screener (Automation Workflow), POC Real-Time Voice Bot (TTS/STT & RAG), POC Chat with Structured Data AI (Agentic), Income Proxy ML Model, Fine-Tuned ASR Model Self-Hosting, Social Listening for Reputation Monitoring Governance, Risk & Responsible AI • Collaborated with the Governance Office to define and implement AI risk management, responsible AI, and AI security policies with technical integration into platforms and workflows. AI Literacy & Organizational Enablement • Drove AI literacy and adoption through internal hackathon coaching, knowledge sharing, and consulting on AI workflow integration across business units. Leadership & Team Management • Leada team of AI/ML engineers and scientists, driving strategy, execution, and technical mentorship across AI initiatives.

Leadership & Team Management • Led Data Science and Machine Learning Engineering teams • Founded MLOps team and operational models to support Data & ML workloads • Lead collaboration between data scientists to streamline transition from POC/experimental state to production ML Platform & Operations • Developed machine learning platform capabilities and technology stack to enable scalable ML solutions • Developed ML drift monitoring tools for distributed system (spark) • Researched and implemented company-wide SDLC standards for Data & ML, including flexible frameworks bridging development, production, and operations activities Advanced ML Solutions & Data Curation • Developed Customer 360 (Data Curations) architecture, including data layer & medallion, ensuring seamless ML compatibility • Developed ML engine on top of Customer 360/Data Curations product to capture key data signals (auto-labeling, categorization, signal profiling, spike triggering) • Prototyped Customer ID solution by consolidating multiple data sources and leveraging network graph algorithms to construct and approximate entity linkage across 5 company subsidiaries

Data Sharing & Customer Matching • Designed and implemented a customer data matching ML algorithm & architecture enabling cross-industry analytics between telco and adjacent sectors (retail, e-commerce, digital finance) across 50+ subsidiaries within CP Group and True Corp. Real-Time Browsing Journey & Triggering • Researched and developed large-scale internet browsing journey extraction to support real-time marketing triggers, complementing micro-segmentation and improving campaign optimization. Insight Quality & Business Impact • Led data science–driven data quality assurance using survey design and statistical methodologies, structured around business impact, modeling assumptions, and data readiness. • Delivered the department’s highest-revenue analytics product, serving multiple external enterprise clients across diverse industries. Customer Lifetime Management • Partnered with a dedicated business unit to drive measurable business value through Customer Lifetime Management and continuous expansion of customer profile depth. ML Engineering & Operations • Contributed to the establishment of a central ML Engineering and MLOps function, defining standards and workflows for hybrid cloud and on-prem data platforms. • Collaborated with platform and data modeling teams to research, develop, and integrate feature store capabilities. Leadership • Led a team of 6 data scientists and machine learning engineers within the Audience Data Science function. • Supervised an undergraduate trainee and supported talent development within the team.

Insight for Customer Lifetime Management • Designed, productionized, and operated a customer segmentation product layered on top of Customer 360, delivering a reusable application framework for lifecycle-driven analytics. • Built segmentation models combining customer life-stage and lifestyle dimensions across time, frequency, and intensity, resulting in 200+ predefined, ready-to-use segments broadly adopted across business units for marketing activation. Micro & Similar Segmentation • Researched and implemented a semi-supervised segmentation approach to generate personalized customer segments from web browsing traversal patterns enriched with behavioral and profile attributes. • Significantly improved campaign lead precision while uncovering previously unknown and adjacent customer segments, serving as a complementary, on-demand segmentation layer alongside Customer 360. COVID-19 Analytics & Public Sector Collaboration • Contributed to COVID-19 risk score estimation using cellular network geolocation and mobility data, in collaboration with Thailand’s Department of Disease Control, to support national pandemic response strategies. Customer 360 – Platform Enablement & Adoption • Enabled enterprise adoption of Customer 360 by aligning platform capabilities with sales, business solution, and campaign analytics requirements, and translating them into clear integration and delivery roadmaps. • Initiated and prototyped core data management and catalog capabilities to support scalable data discovery, ownership, and governance. Leadership • Mentored a team of 5 (data scientists and machine learning engineers) within the Data Science function. • Supervised an undergraduate trainee and provided ongoing technical consultation and knowledge sharing across the organization.

Enterprise Customer Intelligence Data Product • Architected and delivered a production-grade Customer Data Product from inception, integrating scalable ML-driven data pipelines, automated workflows, and governance controls on an on-prem distributed environment. • Transformed heterogeneous digital exhaust (web behavior, geolocation, financial signals, profiles, and service usage) into high-impact customer intelligence for 30.6M subscribers, materializing 400+ structured customer attributes. Large-Scale Analytics & Segmentation • Designed advanced customer segmentation and profiling frameworks using supervised and unsupervised learning, aligned with real-world business constraints at telecom scale. • Enabled downstream use cases across marketing, personalization, and analytics through robust and explainable feature representations. Distributed Data Science Enablement • Created and maintained an internal Spark-based data science framework adopted across core products and exploratory analytics. • Implemented scalable utilities for data profiling, feature selection, multidimensional sampling, Spark ML pipelines, dataset compaction, Hive table lifecycle management, and HDFS operations, significantly improving developer velocity and team autonomy. Leadership & Operations • Established and led an Audience Data Science function focused on customer insight generation and profiling. • Mentored junior talent and advised cross-functional data science teams on architecture, modeling, and scalability challenges. • Adapted agile practices to data science R&D to balance experimentation with reliable delivery. Recognition • Awarded Best Employee Dedication (2019).

Data Pipeline Automation (project base for Bangkok Expressway and Metro Co., Ltd. – Public Transport) • Designed and built automation framework for data & machine learning pipelines of Bangkok subway train on top of GCP.

Product Recommendation System (project base for Sansiri PLC – Real Estate) • Designed and developed hybrid recommendation engine for real estate products. The system serves through RESTful API and able adapt itself to regularly based on its accuracy performance.

Churn Prediction System (project base for Sansiri PLC – Real estate) • Developed and deployed prediction model by applying data mining and machine learning techniques to predict customer booking cancellation and installment plans failure. • Developed ML pipeline and deployment package for RESTful API to serve on-demand prediction approach. It also supports automatic model update. • Visualized profit gain and loss through simulation on dashboard based on prediction result for decision making.

• Taught and developed data science training programs

• Taught and developed data science training programs focusing on customer segmentation and lifetime value analysis for retail business

• Taught and developed data science training programs focusing on text mining for social listening business

Hard Disk Drive Failure Prediction System • Developed and deployed machine learning model and framework to detect and grade potential hard disk drive deflection based on head media parameters to shorten production line lead time and reduce waste in the process. The models coped with extremely imbalanced and overlapping class data and thousands of variables. • Developed near real-time ML pipeline to integrate to sensory data processor machine in HDD production line • Deployed the model using champion-vs-challengers strategy • Developed performance result monitoring and evaluation system to profile model robustness in production incorporation with statistical thresholds for automatic model update. The framework design was quite flexible so that models can be scaled up, managed, monitored easily.

Thesis Keywords: Class Imbalanced, Classification, Hybrid-Sampling, Algorithm, Predictive