Education

NATIONAL UNIVERSITY OF SINGAPORE (NUS)
Master of Computing in Artificial Intelligence

• Cloud Computing (AWS, Azure, GCP, VPCs, K8s, MapReduce-based distributed processing)
• Focus on high-performance networking, probabilistic data structures, and cloud platforms for scalable, low-latency, data-intensive systems

NATIONAL UNIVERSITY OF SINGAPORE (NUS)
Bachelor of Business Administration, Honours (Distinction)

• Specialization in Business Analytics
• Second Major in Data Science and Analytics

Work Experience

PKF-CAP LLP, SINGAPORE

• Built a self-service analytics platform with Role-Based Access Control and n8n automated workflows to streamline operational processes (e.g. CaseWare access, charge code access, Quality Control forms); reduced end-to-end workflow lifecycles by 70%
• Integrated operational data sheets and upstream sources to analytics dashboards via n8n webhooks to achieve pseudo real-time data updates and analytics for partners’ review and decision-making, reducing data aggregation workloads by 80%

AMMO AI, LONDON

• Led full-stack development for Fakers Labs, an experimental AI content lab, shipping features such as AI-generated podcasts that orchestrated multi-step LLM pipelines (outline → script → style transfer → TTS) on top of existing content, increasing user engagement by 30%
• Developed a Twitter AI agent for context-aware responses and analytics dashboard to run A/B tests on prompts and measure engagement (sentiment, click-through), enabling data-driven iteration on model prompts and content strategies, supporting 1.5k daily interactions

MODULAR ASSET MANAGEMENT, SINGAPORE

• Streamlined operational workflow for day-to-day trade booking with full-stack web dashboard to display centralized birds-eye tabulated view of disputed trade info from disparate FXPB and MTM platforms, supporting sub-10 seconds pseudo real-time updates of data, and searching, filtering, sorting, column drag&drop, conditional highlighting

PFK-CAP LLP

• Directed implementation of ‘Potato Audit’, an automated software using Flask as backend, Selenium for streamlining repetitive tasks, and React for developing a user-friendly front-end interface accessible by non-technical audit staff
• Reduced financial announcements and statutory audit processing time from days to sub-10 minutes and semi-automated invoice data extraction of similar formats from manual reviews with 98% accuracy, handling ~28 invoices/min
• Engineered a new systematic staff booking system to minimize booking errors and enhance resource management efficiency
• Received commendation for developing innovative automation solutions - View Testimonial

JASPER VAULT TECHNOLOGY PTE LTD, SINGAPORE

• Developed and optimized dashboards analyzing Ethereum (ETH) Staking, Smart Money holdings and crypto market trends using Dune SQL; presented in weekly investment meeting for fund management
• Streamlined data retrieval and analysis workflows by integrating various API providers using Python/Power Query
• Researched and backtested AI-powered algorithmic trading with optimized parameters on TradingView

Personal Projects

CTO – CHIEF TALENT OFFICER

• Built an AI-powered career platform that ingests a candidate’s GitHub, LinkedIn, and project docs to produce a unified enriched profile, then uses LLMs to generate async tailored resumes, cover letters, and outreach messages per job description within 30 seconds
• Implemented RAG over user-provided sources, so the LLM grounds its responses in actual documents and project history instead of a single compressed resume, reducing hallucination by 20%
• Integrated Apify-based flows to search for HR / hiring-manager contacts and link them with LLM-generated personalized outreach content, closing the loop from job description → profile → contact, reducing manual search workload by 30%
• Designed the system so it can be extended into a conversational analytics interface over a user’s career data (e.g. “What are my top 3 projects relevant to this JD?”)
• See cto.shiyao.dev and GitHub

N8N HEADHUNTER WORKFLOW

• Developed workflows that take natural language query → parse into structured filters → search and enrich profiles → top 20 candidates in ave. 2min
• Deployed search-first approach using Google (via Serper) + LinkedIn dorking + Apify enrichment for fresher, lower-cost data ($0.05 per profile)
• Configured LangChain-based code node to stabilise tool-calling, ensuring the agent loops until the target number of candidates is found and avoiding random early termination, improving accuracy by 20%
• Logged all agent actions (searches, validations, accepts/rejects) to Supabase and exposed them via webhook for live progress UI, solving the “why is nothing happening?” problem and making the pipeline 90% auditable and explainable
• Connected to a people dataset (200M+ entries) for fast filter-based retrieval (5–10 seconds per batch), while parallel Yahoo/Google search improves coverage for profiles not in the dataset; designed for future semantic search / RAG over stored profiles

SG JOBS AGGREGATOR

• Built 20+ n8n workflows to scrape Singapore companies’ official career sites, normalise heterogeneous APIs/HTML, and aggregate 4,000+ listings in about 20 minutes (title, apply URL, metadata) into a unified dataset
• Implemented an end-to-end mini data platform: information gathering → cleaning → DB insert → JSON and index.html generation → deployment to GitHub Pages, orchestrated as scheduled daily flows
• See sg-jobs and GitHub

Co-Curricular Activities

COMLINK BEFRIENDERS

• Engaged with local families for monthly visits and activities

Skills and Interests

Languages & Databases
Python, Java (OOP & data structures – NUS CS2040), C, JavaScript, TypeScript, SQL, MongoDB, PostgreSQL, SQLite

LLMs, AI & Automation
LangChain, CrewAI, RAG, OpenAI, Dify, TTS, Whisper, OCR, n8n, Make.com, Supabase

Data Platforms, Web & Services
React, Next.js, Node/Express, Flask, REST APIs, Selenium
Experience building internal analytics dashboards and data-collection services over financial, audit, and job-market data

Cloud & Infrastructure
Azure (AI Foundry, Document Intelligence, VMs, NSGs), AWS (EC2, VPC, S3), Digital Ocean, Docker, Traefik
Familiarity with distributed systems concepts (MapReduce, cloud-native architectures) from NUS Cloud Computing coursework