B4se

N0de

Designed and implemented a REST API for processing requests from Telegram and Task Tracker. Optimized routing, middleware, and database interactions for high-load scenarios. Designed the database schema and implemented transactional CRUD operations using SQLAlchemy with async support. Built and optimized Docker images using multi-stage builds and improved docker-compose setup for scalable deployment. Integrated LLM-based logic for ticket processing, including automatic request categorization. Implemented caching and rate limiting for GPT API. Key achievements: Built an automatic token refresh system eliminating manual deployment steps and improving reliability. Implemented health checks, logging, and Telegram alerts for continuous monitoring. Automated ticket classification using GPT, significantly reducing manual operator workload.

Developed a monitoring dashboard displaying real-time status of controllers, sensors, MQTT broker, databases, and Telegram bots, using static rendering for performance and availability. Analyzed and optimized system configurations, improving data accuracy and reliability. Implemented ADC serial number tracking to prevent configuration loss during controller updates. Designed an incident notification escalation system and built weekly/monthly reporting for scheduled inspections. Key achievements: Co-developed a physical health-check device with embedded software. Optimized database architecture, reducing storage size by 50% and improving query performance by ~1.5×. Built an automated sensor configuration and calibration system based on historical data.