Nemanja Adamov

Sports betting platform · Python · PostgreSQL · RabbitMQ · Redis · WebSockets

Backend Engineer (2015) → Backend Team Lead (2019) → Software Architect / Senior Full Stack Engineer

• Owned the end-to-end architecture and infrastructure of the betting platform: database layer, application servers, web layer and deployment topology.
• Architected and evolved an event-driven betting platform serving 5,000+ active users, with 30+ Python queue consumer daemons (Redis, RabbitMQ) for live odds, odds generation and ticket pipelines.
• Designed and implemented a microservice-oriented architecture with clear service boundaries for sportsbook, lottery, slots, risk management and cashout flows.
• Implemented database read/write splitting and server load balancing to keep the platform stable under peak betting loads.
• Built probability models and odds engines in Python, optimising performance for latency-sensitive odds calculation.
• Led and mentored a backend team of 8 engineers: architecture definition, code review, technical leadership and quality standards.
• Drove architectural decisions and hands-on development across the stack, including LLM-based automation in key business workflows.

• Founded and built a taxi dispatch platform end-to-end — from concept and system design to production deployment.
• Sold the platform to another company; continue to maintain and evolve the system in production use.
• Designed a Python microservice backend and REST APIs, plus driver tablet and passenger mobile apps (Android/iOS) and JavaScript back office.
• Designed and implemented an in-vehicle IoT device (GPS tracking, panic button) integrated with the driver tablet and dispatch backend.

Delivered two internal business applications: Python backends, JavaScript frontends.

Maintenance and feature development of a sports betting platform, including integrations with payment peripherals (coin/bill acceptors, POS printers).

Own product · Python microservice framework · FastAPI · async SQLAlchemy · PostgreSQL · Redis · Celery

• High-performance framework for building cloud software: from a single YAML spec it generates models, schemas, APIs, CRUD services, admin back office and documentation.
• Infrastructure (auth, multi-tenancy, RBAC/ABAC, real-time events, background workers) comes out of the box so development can focus directly on business logic.
• Designed for AI-assisted development: agent manifests describe the architecture so LLM-based tools can safely generate and extend services in a consistent way.
• Extended xapi with an AI agent that lets me describe new services in natural language; the agent generates the YAML spec, after which the framework produces the full service stack automatically.

Own product · built on xapi · web + mobile · event-driven · Docker · xtms.app

• Multi-tenant transport management system covering fleet, drivers, load dispatch, transactions and invoicing, with a workflow (flows) engine and full internationalisation.
• Includes a dedicated AI microservice for load-related automation (e.g. suggestions and decision support for dispatch and planning).
• Each microservice is generated from declarative schema definitions; the platform is fully event-driven and Dockerised across development, staging and production.
• Built both the web application and the companion mobile app, covering the entire operation end-to-end — from back office staff to drivers on the road.

Own product · asyncio · PostgreSQL RLS · Docker

• Designed and built in Python: an asyncio daemon that speaks the SAS slot-machine protocol directly with gaming machines.
• Multi-tenant cloud aggregator with per-location administration for casino floors.
• PostgreSQL row-level security for strict tenant isolation, field-level encryption for player data, and a hardened Docker deployment for production environments.

• Focus areas: databases, network programming, software engineering and cyber security.

• Integrating LLM APIs (Anthropic, OpenAI) into production Python services — automation, internal tooling and document/data workflows.

• Running self-hosted AI infrastructure: open-weight models on own servers plus an n8n-style workflow orchestration system powering automations on current projects.

• Designing agentic workflows: multi-agent pipelines for code generation, review and verification, used daily in real product delivery.