Programming

Problem: Large volumes of legacy paper documents containing sensitive clinical data needed to be digitized and made searchable within the EMR system, but manual data entry was prohibitively labor-intensive.

Solution: AI-powered document processing pipeline using OCR and open-source LLMs to extract, structure, and format legacy clinical data for EMR integration.

Problem: Manual scheduling causing double-bookings and inefficient time slot utilization.

Solution: Automated scheduling system with conflict detection and optimization algorithms.

Problem: Difficulty visualizing patient progress over time using scattered paper records.

Solution: Data dashboard for tracking and visualizing rehabilitation metrics and outcomes.

Problem: Repetitive documentation tasks consuming significant clinical time.

Solution: Template-based documentation tool with smart text expansion and common phrase library.

Problem: Patients forgetting home exercise programs after discharge.

Solution: Curated video library with QR-code linked exercise sheets for patient take-home.

Problem: Patient data scattered across multiple clinical systems requiring manual cross-referencing.

Solution: Read-only data aggregation tool for unified patient view during assessments.

Problem: Creating monthly caseload reports manually was tedious and error-prone.

Solution: Automated script for generating standardized monthly statistics and reports.

Problem: During the pandemic, meat shortages and price volatility made it difficult for consumers to plan purchases and find available products at reasonable prices.

Solution: Automated web scraping and data analysis tool to monitor meat availability and pricing trends across GTA wholesalers.

Problem: Need to send hundreds of customized emails through ProtonMail to connect parents of children with autism to Ontario Autism Program services, but ProtonMail lacked traditional API access for programmatic sending.

Solution: Selenium-based automation tool with dynamic templating system to send personalized emails at scale through ProtonMail's secure interface.

Problem: Need for at-a-glance visibility into server status, outages, and key information without requiring a computer or phone.

Solution: Custom IoT LED matrix display powered by ESP32 that pulls real-time data from local Express.js endpoints.

Problem: During the GPU shortage, finding available graphics cards required constant manual checking across multiple retailers.

Solution: Automated web crawler that monitors retailer websites for GPU availability and sends alerts when stock is detected.

Problem: Traditional monitoring solutions fail when WiFi and cellular networks are compromised during outages, leaving administrators blind to server status during critical events.

Solution: Battery-backed LoRaWAN transmitter system that operates independently of internet infrastructure to monitor and log server uptime.