Portfolio — 2026
Singapore · SMU Finance & Strategy ’27
Analyst by training,builder by instinct.
Finance & Strategy undergraduate at Singapore Management University and Distribution Analyst intern at Conduit Group. I do the work first — fund due diligence, LP analytics, valuation — then build the systems that make it reproducible: 34 tearsheets across 8 alternative-asset sectors, a 5,146-contact LP engine, and an automated equity-research desk that briefs me every morning.
Flagship work — 01
Fund distribution · Due diligence · Investor relations
The fund due-diligence & distribution pipeline
At Conduit Group I turned raw alternative-asset fund PDFs into rated, documented shelf decisions — replacing manual triage for the investor-relations team.
From PDF to verdict
LLM extraction pulls strategy, terms, track record, and risk data out of raw fund documents; a 7-factor scoring framework turns it into a 14-section due-diligence workbook with a composite rating and an A/B/C/Pass shelf verdict the IR team can defend.
Coverage at intern speed
Processed ~17 fund managers across 8 alternative-asset sectors — hedge funds, private equity, real estate, VC, quant, private debt, digital assets — shipping 34 manager- and sub-fund-level tearsheets plus outbound IR emails and investment commentary.
The LP engine
Rebuilt the firm's 5,146-contact LP database as v2.0: LLM-driven matching ranks every contact into Tier 1/2/3 prospect bands for each new fund, and the whole thing is queryable in natural language instead of by manual row-scanning.
Workflow leverage
Shipped the supporting tools the team now runs on — automated HTML email broadcasts through Monday CRM, LLM-driven web scraping to enrich the company database, and scripted automations across recurring processes.
- 14-section due-diligence workbook with composite rating and A/B/C/Pass shelf verdict
- 34 tearsheets across ~17 managers and 8 alternative-asset sectors
- 5,146-contact LP database with Tier 1/2/3 prospect matching, queryable in natural language
Case studies — 02–08
Selected work
Hover to open
An automated equity-research desk
Personal research system — nights and weekends
To learn institutional research discipline properly, I built the whole stack: role-based AI agents (analyst, PM, risk, macro, scout) write dated research memos on a schedule, a valuation engine runs DCF and comparables on every name, and market data is grounded in WRDS — Wharton's institutional research database — with cross-source integrity checks.
Equity research · Valuation · Financial data
Worker Master — running a real agency on AI ops
Smile Employment Agency — my own licensed company
The operations platform behind my employment agency: a spreadsheet-backed workforce system with a mobile app, per-company quota calculators, and AI-driven compliance audits — supporting 50+ work-pass placements and 20+ corporate clients.
Business operations · Compliance automation · Founder
The Post Room — an AI inbox concierge
Email + learning system over two Gmail accounts
A dependency-free Python server plus headless LLM workers that read two Gmail accounts and produce a daily morning brief: grouped newsletter digests, extracted deadlines, and a finance-news desk with a spaced-repetition concept ledger that turns reading into retained knowledge.
Agentic AI · Systems design · Finance learning
Telegram concierge & mission-control board
Personal finance + operations assistant
A Telegram bot that grew into a personal operations hub: expense tracking with receipt OCR, bank-statement parsing, and a worker-edit queue for my agency — audited by a mission-control dashboard that health-checks every feature.
Agentic AI · Product thinking · Automation
Idea Scout & automated research pipelines
Systematic learning and idea generation
A weekly multi-source scout that scans GitHub, Hacker News, Reddit, and YouTube, scores every candidate against a written interests profile, and publishes a ranked digest — alongside a deep-research harness that fans out parallel agents to produce cited, fact-checked reports.
Research automation · Multi-agent systems
The Desk — a personal cloud of ~60 agents
The umbrella system tying everything together
Every system I build ships to one self-hosted hub: a dashboard of dashboards with a registry of ~60 agents and scheduled automations, health watchdogs driving live status lights, 30+ custom AI skills, and usage telemetry.
Infrastructure · Agent operations · Self-hosted
House Finder — an autonomous housing agent
Finding my own exchange-semester flat in Dublin
An end-to-end autonomous agent for a real personal problem: it scans Dublin housing sources, scores fit against my constraints, sends enquiries as me by email and browser form-fill, and answers landlord replies with an LLM reply-brain — escalating only high-stakes messages for one-click approval.
Autonomous agents · Real-world task automation
Also built
The wider workshop
About
I can't leave a manual process alone.
At my internship I did the fund triage by hand once — then built the pipeline that turns a raw fund PDF into a rated, documented shelf decision. That pattern runs through everything here. I founded a licensed employment agency the same year I started university, then automated its compliance paperwork. I wanted to genuinely understand markets, so I built an equity-research desk that produces valuation work and morning briefs on a schedule. I wanted to learn faster, so I built a system that turns what I read into spaced-repetition reviews.
What drives me is straightforward: I'm obsessive about getting better — at analysis, at building, at the craft of whatever is in front of me — and I don't wait to be told to start. Every system on this page exists because I refused to do the same work twice. Put me on a team with real problems and I bring the same habit: learn it properly, do it well, then make it faster for everyone.
Experience
Where I’ve worked
Conduit Group
Distribution Analyst Intern
Jan 2026 — Present
- Designed a distribution pipeline pairing LLM extraction with a 7-factor scoring framework, converting raw fund PDFs into a 14-section workbook with composite rating, A/B/C/Pass shelf verdict, and factsheet — replacing manual triage with reproducible insights for the IR team.
- Processed ~17 fund managers across 8 alternative-asset sectors, shipping 34 manager- and sub-fund-level tearsheets and drafting outbound IR emails and investment commentary.
- Shipped v2.0 of the firm's 5,146-contact LP database — LLM-driven matching ranks contacts into Tier 1/2/3 prospect bands per new fund, queryable in natural language.
- Built supporting tools the team now runs on: automated HTML email broadcasts via Monday, LLM-driven web scraping for database enrichment, and scripted automations across recurring processes.
Smile Employment Agency
Founder
Apr 2024 — Present
- Founded and manage a licensed Singapore employment agency (UEN 24C2161) specialising in service-sector manpower placements.
- Achieved 50+ S Pass and Work Permit placements within the first 8 months, outperforming industry averages through workflow automation, document standardisation, and proactive compliance tracking.
- Maintain a client base of 20+ corporate partners at a 90% retention rate; every placement completed within 10 working days.
Commerce.sg
Business Development Executive
May 2024 — Aug 2025
- Identified B2B prospects and partnerships in the social-commerce space; campaigns generated US$100,000+ in attributed client revenue.
- Tailored influencer and content strategies to client-specific audiences to boost reach and brand recall.
New V Spa
Marketing Operations Intern
Sep 2022 — Feb 2023
- Owned social-media strategy across Facebook, Instagram, and TikTok; produced short-form videos reaching 100K+ organic views.
Toolbox
What I work with
Finance
- Valuation (DCF, comps)
- Fund due diligence & tearsheets
- LP prospect analytics
- Equity research workflow
- Portfolio & risk memos
- Backtesting discipline
- Excel (Solver, PivotTables)
AI systems
- Claude Code & headless agents
- Multi-agent orchestration
- MCP servers (built my own)
- Custom skill design (30+)
- LLM extraction & scoring pipelines
- Prompt & context engineering
Engineering
- Python
- JavaScript/Node
- Flask & stdlib servers
- Google Sheets/Gmail APIs
- Playwright automation
- Docker, nginx, systemd/launchd
- GitHub Actions CI/CD
- Self-hosted VPS ops
Business
- Founded & run a licensed agency
- Client management (20+ B2B, 90% retention)
- Regulatory compliance ops
- Monday CRM
- Digital marketing
- Mandarin Chinese (fluent)
Education & community
Background
Singapore Management University
BBM, Double Major in Finance and Strategy · GPA 3.7/4.0
Aug 2024 — Aug 2027
Ngee Ann Polytechnic
Diploma in International Trade and Business
Jan 2019 — Dec 2022
SMU Project Smile Manila XV
Project Lead
Jan 2024 — Present
- Lead a 24-member SMU team for an overseas community service project with CRIBS Foundation (Philippines), shifting the project from short-term aid to long-term sustainability through financial-literacy workshops and corporate-partnership training.
Contact
Let’s talk.
Hiring for asset management, fund distribution, or strategy? I bring the analyst toolkit and the AI leverage to go with it.
cy.wang.2024@business.smu.edu.sg