Singapore / SF · Agent Architect

Celeste Ang.

Data Scientist ML Engineer DevRel Agent Architect

9 years of experience in data and AI across fintech, enterprise tech, and the agentic frontier. Currently Head of DevRel & Forward Deployed Engineering at Virtuals Protocol — building AI agents, architecting frameworks, and growing the developer ecosystem.

Celeste Ang

Write


Currently building

Head of DevRel & Forward Deployed Engineering @ Virtuals Protocol

Leading a technical team building AI products — companion agents, evaluator agents, and adtech agents seeding an open agent marketplace. Architecting the frameworks that other builders build on. Speaking at ETH Denver, Coinbase hackathons, and across the global agent ecosystem.


Embedded AI advisory, from architecture to implementation

I work with startups and enterprises as an embedded, forward-deployed engineer: close enough to understand the workflow, experienced enough to architect the system, and hands-on enough to implement it end-to-end. I build practical AI agents and automations that remove manual work and help teams move faster.

With experience spanning agentic products, workflow automation, and developer enablement, I thrive in helping developers and teams turn messy processes into useful agents, workflows, and internal accelerators, either as a solo developer or within a lean team.

Workflow automation Agent systems AI enablement
Get in touch

Experience

2025 — Present
Virtuals Protocol
Head of DevRel & Forward Deployed Engineering

Building and maintaining AI products. Architecting agent frameworks and seeding the marketplace with companion, evaluator, and adtech agents. Leading developer relations and mentoring engineers across the ecosystem.

2025 — Present
AI Advisory Practice
Embedded AI Advisor & Forward-Deployed Engineer

Advising and implementing end-to-end AI agent systems for startups and enterprises, from workflow diagnosis and solution architecture to production-ready builds. Works solo or in lean teams to automate manual processes, improve internal operations, and make AI adoption practical for day-to-day work.

2024 — 2025
Heicoders Academy
Data Science Instructor

Taught and supported data science and analytics classes covering Python, SQL, Tableau, scikit-learn, and applied analytics workflows. Helped learners make technical concepts practical through hands-on exercises and real-world tooling.

2024 — 2025
Oliver Wyman
Lead Data Scientist / Engineer

Designed the data model for Thailand's largest bank, optimised for a lakehouse architecture on Databricks. Databricks partner champion.

2022 — 2024
Sourced Group
Consultant, Data Analytics & ML

Designed Azure data mesh platforms and AWS data architectures for a major Japanese telco and Malaysian FSI. Built ML models for HR analytics using Databricks model registry and serving. 11 cloud certifications across Databricks, Azure, and AWS.

2019 — 2022
Grab Holdings
Senior Data Scientist / Engineer

Built data pipelines and datamarts for 5+ SuperApp categories with 9+ external partners. Developed causal inference models and A/B test frameworks to quantify product impact at scale. Built an internal tool to democratise causal inference across the org.


Case Studies


Events


Socials


Skills

Python SQL R PySpark / Spark Databricks AWS Azure Tableau MongoDB Quicksight Agent Frameworks LLMs Causal Inference ML Engineering Lakehouse Architecture Data Mesh DevRel

MSc in Computer Science
Georgia Institute of Technology
2018 — 2020
Specialization in Machine Learning
BSc in Business Analytics
National University of Singapore
2014 — 2017
NUS Overseas College Scholar
Exchange Stanford University

2019 2nd Place — Rolls-Royce Data Challenge · ML & Monte Carlo simulation for flight delay prediction
2017 Certificate of Distinction, Analytics Techniques · Certificate of Merit, Data Science — NUS Computing
2016 NUS Overseas College Scholarship — Year-long startup internship in Silicon Valley with coursework at Stanford University
2016 Finalist — NUS School of Computing Innovation Prize · Ocubulum: automated research analytics platform

Resume

Your browser can't display the PDF inline — download it here.