Hi, I'm Assylbek

• Built ProfSidekick, a full-stack AI teaching assistant platform that transforms static presentations intovoice-interactive lessons with real-time AI tutoring and slide control..

• Engineered scalable backend (FastAPI, PostgreSQL, Redis) with 25+ REST APIs, 6+ database models, deployed via Docker/Fly.io.

• Integrated OpenAI Realtime, Vision, and Function Calling APIs, enabling live voice Q&A, automated slide explanations, and AI-driven slide navigation.

• Implemented real-time communication via WebRTC and custom event streaming, enabling live voice processing, AI guidance, and session syncing across users.

• Engineered an LLM-powered assistant, CTOgram, for mobile and Telegram, streamlining car service and part requests; achieved a user satisfaction score of 4.8 out of 5 based on post-interaction surveys.

• Developed FastMCP, a custom Remote Model Context Protocol (MCP) server integrated with the Responses API, enabling real-time tool execution and structured multi-turn conversations.

• Improved UX with dynamic UI elements (buttons, keyboards) based on MCP outputs, e.g., showing user’s cars via phone lookup for instant selection.

• Integrated PostgreSQL with Redis caching layer, resulting in a 30% reduction in database query times; streamlined data access for 4 core services utilized by 5+ engineers across the organization.

• Led integration for Pizza Express and 20+ restaurants across 4 regions, enabling 50K+ monthly transactions via Python microservices, REST APIs, MSSQL, and Redis.

• Delivered integrations for 5+ global POS providers, shortening onboarding timelines by 40%, by standardizing API development, architecture design, and effective stakeholder collaboration.

• Implemented LLM-based categorization system, achieving 94%+ accuracy in automated item classification, improving restaurant analytics and operational insights.

• Improved system reliability and decreased troubleshooting time by 40% by implementing distributed tracing (Jaeger) and real-time monitoring dashboards (Grafana).

• Developed and deployed QR-based menu ordering system for a key POS provider, increasing ordering efficiency by 25% and enhancing customer experience through seamless integration.

• Developed a Chrome extension capturing 2M+ user events for productivity insights in real-time.

• Built secure Flask backend with 99.9% reliability using JWT authentication for real-time event collection.

• Optimized PostgreSQL, decreasing query times by 40% with indexing and partitioning.

• Deployed a scalable Flask application on AWS using EC2, Elastic Load Balancing, Auto Scaling.