Co-led a 3-person AI team developing an automated grading pipeline for business case-study reports, slashing professor grading time from approximately 30 hours to under 8; the system generates structured per-rubric scores and feedback, stored in PostgreSQL, and editable via instructor controls
Designed a modular multi-agent architecture with LangChain, featuring callable Python modules for rubric parsing, case retrieval, and submission analysis; enforced schema-controlled prompts and outputs, with instructor edits supported through a Flask API and React frontend
Built differentiated document pipelines: chunked and indexed case files into FAISS using sentence embeddings for efficient retrieval-augmented prompts, while preserving full student submissions for holistic evaluation—reducing token usage by approximately 60% without compromising
Engineered rubric-to-principle conversion logic, transforming criteria into JSON rules; each principle triggers targeted prompts to compare submission arguments against case context, producing per-principle scores and justifications for faculty review and optional revisions
Teaching Assistant
University of Toronto Education Center
Toronto, Canada
05.2024 - 08.2024
Course: Cloud-Based Data Analytics
Guided 101 students on using Spark on the Databricks and data retrieving and transferring on the Azure cloud platform
Held weekly office hours for 25 ML-Project Teams, resulting in a 5% increase in final project grade
Research Assistant
Rotman School of Management - University of Toronto
Toronto, Canada
09.2022 - 04.2024
Supervised by Professor Kim (Accounting Area)
Automated extraction of director biographies and firm expertise needs from 15,000+ SEC annual & proxy filings covering 1,000+ companies, replacing manual parsing workflows
Designed a Multi-Agent LLM workflow that classified directors’ functional and experiential expertise and matched it to company needs with about 92% precision, cutting manual labeling time by roughly 85%
Delivered a clean longitudinal dataset now used in empirical studies on board-expertise alignment and firm performance
Financial Analyst Intern In Asset-Backed Securities (ABS) Team
CITIC Securities Company
Beijing, China
05.2022 - 08.2022
Gathered and analyzed financial data across 30+ cities using the firm’s bond system; created visual reports and decks that enhanced communication and accelerated team review workflows by ~20%
Assisted senior analysts in conducting in-depth analysis of 34 financial metrics, applying clustering techniques (DBSCAN) to predict ABS distribution patterns, aiding in strategic decision-making
Built automated web-scraping scripts from Python to extract bidding scope data from 24 platforms and 19 companies, reducing manual research time and improving data freshness for the team
Education
Master of Engineering - Electrical and Computer Engineering
Bank Transaction System (Ledger Simulator), Modelled core retail-bank workflows in a React + TypeScript Frontend and Express/PostgreSQL Backend: secure auth, deposits/withdrawals, intra-account transfers, and paginated ledger views, Containerised the stack (Docker + Compose) with one-command spin-up, giving Professor a Two-Minute interactive Demo and holding team meeting weekly demonstrating best-practice CI/CD for finance-grade software.
Timeline
AI Engineer
Alagora (EdTech Start-up)
05.2025 - Current
Teaching Assistant
University of Toronto Education Center
05.2024 - 08.2024
Research Assistant
Rotman School of Management - University of Toronto
09.2022 - 04.2024
Financial Analyst Intern In Asset-Backed Securities (ABS) Team
CITIC Securities Company
05.2022 - 08.2022
Master of Engineering - Electrical and Computer Engineering
Chief of Strategy and Marketing Operations at HERO [EdTech Start-up, Online Tutoring Platform]Chief of Strategy and Marketing Operations at HERO [EdTech Start-up, Online Tutoring Platform]