🥦

Zhang Jiayi

Y2 CS Student • Passionate Problem Solver • Product-minded Systems Builder

I turn complex problems into well-designed systems, from machine-learning tools to automation and full-stack applications, with a focus on clarity, efficiency, and real-world impact.

Photo of Jiayi

Study in NUS

  • Year 2 Computer Science with 2nd Major in Management
  • Relevant modules: Data Structures and Algorithms, Software Engineering, Introduction to AI and Machine Learning
  • Co-curricular experience: Director of Student Engagement, NUS Students' Computing Club, Singapore

Product Experience

2nd Place • PayPal × NUS Product Club Product Launchpad

Jan 2026 - Feb 2026 • Singapore
  • Selected for a 6-week intensive PM workshop and concurrent case competition at the PayPal Innovation Lab, moving from zero-to-one product management through weekly in-person sessions.
  • Led the discovery phase using the GAP framework to identify that 70% of financial institutions lose clients during onboarding; Validated the "Momentum Gap" for SMEs through qualitative research and market trend analysis.
  • Architected an AI-Adaptive Parallel Onboarding System featuring a Dynamic Engagement Engine (Yield & Cost of Wait Simulators) and a Risk & Revenue Orchestration Engine to prioritize high-value activation.
  • Utilized JTBD to define user goals and applied MoSCoW prioritization to scope an MVP that balances regulatory compliance with immediate "Investment Sandbox" engagement. Designed a three-phase Beachhead GTM strategy and implemented Shadow Mode and Blind QA protocols to mitigate AI automation bias and ensure global model transferability.

Product Specialist Intern • Fly Fairly

Jan 2026 - Present • Singapore
  • Working on the end-to-end development of a high-impact Operational Reporting Suite using Next.js, managing the full product lifecycle from stakeholder discovery and scoping to build and validation.
  • Engineering data reports for engineering, product, marketing and operation departments within the back office, managing the technical trade-off between real-time data freshness and query performance by defining precise metrics and data sources.
  • Focusing on translating complex data sources into actionable metrics and filters to support daily executive decision-making.

Consulting Experience

1st Runner Up (100+ Teams) • SGX-NUS Sustainability Case Competition

Jan 2026 - Feb 2026 • Singapore
  • Spearheaded product discovery to identify the "Complexity Wall" for Gen Z investors and architected a DLT-based fictionalisation strategy that reduced institutional entry barriers by 5,000x ($250k to $50).
  • Designed an API-driven framework to normalise complex ESGenome institutional data into relatable retail impact metrics, bridging the technical disconnect for digital-native users.

Part Time Assistant • Bain & Company

Dec 2025 • Chongqing China
  • Selected to support a private equity due diligence assessing a family planning products brand as store visit temp.
  • Conducted primary market research across multiple offline channels by developing a structured data summary framework using standardised Excel templates to capture and organise critical product attributes, including pricing, shelf placement, and availability

Selected Projects

Facilities Booking Telebot

Built a Telegram-based booking system with real-time availability checks, conflict detection, and approval notifications.

Facilities Booking Telebot screenshot

SoCTAssist

SoCTAssist is a desktop app designed specifically to help Teaching Assistants manage their students' information, homework, attendance, and consultation sessions more efficiently.

SoCTAssist

The Travel Parrot

Used GPT to label 10,000+ raw restaurant reviews and trained a Random Forest classifier with engineered NLP features to filter low-quality reviews.

Travel Parrot screenshot

shAIpping_List@NUS

Built a full-stack AI shopping list generator using GPT-4o, React, and PostgreSQL with locally curated items near NUS.

shAIpping_List@NUS

Gesture-Controlled Browser Management System

Architecting a real-time gesture recognition system using ML Model MediaPipe, mapping 21-point hand landmarks to specific browser commands such as window tiling and tab consolidation

MediaPipe Model Gesture-Controlled

© Zhang Jiayi