Projects
Real systems I've built and maintain. These are the proof points behind the consulting — not classroom exercises or abandoned prototypes.
Automated Purchase Tracking
Flask · Claude API · Google Cloud Firestore · Gmail API
A pipeline that connects Gmail, Amazon order history, and credit card feeds into a single system. Claude API parses every purchase email automatically, extracts structured data (item, price, category, vendor, date), stores it in Firestore, and exports to Excel.
- 974 items tracked, with purchase history back to 2012
- 96.7% auto-classified with no manual entry
- Runs unattended — no daily interaction required
Why it matters for clients: This is the pattern — find a repetitive data task, connect the systems, let AI handle classification, and measure whether it actually worked. The same approach applies to invoice processing, customer intake, and inventory management.
Read the case study →Document Intake & Search
Python · Tesseract OCR · SQLite · Click CLI · Pillow
Scan a physical document with your phone, and the system auto-processes it: runs OCR, extracts text, files it in a searchable SQLite archive, and categorizes it. Receipts feed into the purchase tracking pipeline; business cards become a searchable contact database.
- Processes scans in under 5 minutes from photo to searchable archive
- CLI interface for power users — no web UI overhead
- Phases 1-6 complete; AI-powered extraction planned
Why it matters for clients: Paper-heavy businesses (legal, medical, trades) lose hours to manual filing. This shows how OCR + structured storage eliminates that bottleneck with off-the-shelf tools.
Study Platform
Flask · Supabase · Jinja2 · JavaScript
A web app built for Golisano Institute students. Practice quizzes with class analytics, lesson notes, formula sheets, and quiz history. Real users, real usage every week.
- Live at study.kevinsykes.ai
- Used weekly by active students during the semester
- Backend on Supabase (free tier), secrets managed via GCP Secret Manager
AI Adoption Research
Streamlit · JSON · Data Analysis
Structured interviews conducted at the weekly Golisano Institute Speaking from Experience series — with guest speakers and recruiting firms who set up booths on campus. Captures AI readiness, tool usage, and workflow adoption patterns across industries.
- Data collected from speakers and industry recruiters each Wednesday
- Cross-industry comparison of AI adoption patterns
- Past speakers include leaders from Paychex, Linden Digital Marketing, and Breathe Yoga & Juice Bar