William Kim
Data/Business Analyst
I turn messy data into clear KPIs, dashboards, and decisions. SQL + Snowflake/Python for prep, Power BI/Streamlit for delivery, and pragmatic AI when it accelerates time-to-insight.
SQL • Snowflake • Power BI • Python (pandas) • KPI Design • Visualization • A/B Testing • Forecasting • ETL • LLM/OCR (practical)
About Me
I’m William Kim — a data-driven builder who delivers analyst outcomes: trustworthy KPIs, clear dashboards, and crisp narratives tied to business goals. I work across the stack when needed (SQL/Python → BI), and I’m comfortable adding AI/LLM pieces where they reduce time-to-insight. Recent work includes a Snowflake + Power BI sales dashboard and KPIFlow AI for one-click KPI views from CSV/SQL.
Projects & Case Studies
Snowflake + Power BI Analytics
End-to-end sales analytics: Snowflake SQL data modeling + Power BI dashboard with KPI cards, trends, top products, and category breakdowns.
Repo includes raw CSVs, SQL DDL/DML, and dashboard screenshots. Demonstrates data modeling, ETL steps, and BI storytelling.
- Snowflake
- SQL
- ETL
- Power BI
- KPIs
- Data Modeling
KPIFlow AI
Self-serve KPI dashboards from CSV/SQL with automated insights. Streamlit front-end with Python (pandas), metric definitions, and exportable visuals.
Automated KPI cards, trend charts, CSV → insights; built for analyst-style storytelling and quick stakeholder share-outs.
- Python
- Pandas
- Streamlit
- Plotly
- CSV→Dashboard
- Insights
SplitChamp AI
AI bill-splitting app with Azure Document Intelligence/Vision + GPT-4o/5 pipelines for receipt parsing and structured expense breakdowns.
Play Store-ready: Expo EAS builds, Pro entitlements via Stripe/RevenueCat, feature flags, in-app legal.
- Expo/React Native
- Python
- FastAPI
- Azure OCR
- GPT-4o/5
- Stripe/RevenueCat
GhostTab AI
Chrome extension with OCR-based page capture, Azure OpenAI summarization/rewriting, adaptive tab limits, Pomodoro timers, and contextual coaching. Includes PyTorch → ONNX classifier (onnxruntime-web).
Published on Chrome Web Store; summarizer/rewriter flows + ONNX classifier live.
- Chrome Extension
- Azure OpenAI
- OCR
- GPT-4/5
- Pomodoro
- PyTorch→ONNX
CJ Reflecta On hold
⏸ On hold — AI-powered journaling app for guided self-reflection and emotional growth.
- React Native
- Expo
- AsyncStorage
- GPT-5
Doyrix On hold
⏸ On hold — GPT-assisted investing research tool for filings and earnings transcripts.
- Next.js
- FastAPI
- Supabase
- Stripe
- GPT-5
Analyst Highlights
- Designed decision-ready KPIs (Total Sales, Orders, AOV) and surfaced top customers/products to guide prioritization.
- Built reusable SQL models and clean BI layers for transparent, auditable metrics.
- Cut analysis time by automating CSV→dashboard flows (KPIFlow AI; Streamlit + pandas).
- Applied Python (pandas) for cleanup/joins and Power BI for executive-ready visuals.
Core Skills
- Data: SQL, Snowflake, Python (pandas)
- BI: Power BI, Streamlit/Plotly
- KPIs: Metric design & governance
- Analysis: A/B testing, forecasting
- Ops: ETL basics, data modeling
- AI add-ons: LLMs/OCR (practical)
Certifications
- Databricks Generative AI Fundamentals
- Databricks AWS Platform Architect
- Azure AI Engineer Associate (in progress)
Contact
- Email: williamcjk11@gmail.com
- LinkedIn: william-c-kim
- GitHub: @willckim