Dmytro – Node.js, Typescript, JavaScript

A decentralized telecom verification network that utilizes a global fleet of mobile devices to generate verifiable network performance proofs. The platform allows users to contribute their device's connectivity data to a decentralized network in exchange for rewards, creating a transparent map of global telecom performance. Key features include a mobile node application with native hardware integration, a web management dashboard for network monitoring, and a Web3-integrated verification system for blockchain-based performance proofs.

• Focused on creating seamless integration between mobile devices and cloud services.
• Built the mobile app and web dashboard on a Supabase backend, optimizing SQL queries and RPCs for efficient handling of large-scale network data.
• Managed real-time database subscriptions to ensure up-to-date information across clients.
• Developed custom Native Modules in Kotlin and Swift to capture detailed network metrics from devices. D
• elivered core features including the chat system and push notifications, while automating deployment pipelines with EAS and GitHub Actions for reliable and fast releases.

A B2B SaaS platform that streamlines software and partner discovery through AI-powered capabilities and advanced search mechanisms. It also functions as a lead generation platform, enabling software vendors and partners to identify and connect with relevant customers.

• Led a team of 6 people and owned end-to-end architecture and delivery of the platform across frontend, backend, and infrastructure.
• Designed and implemented the dual search architecture (fuzzy + exact search and AI-driven semantic search), including vector search integration with Qdrant and AI pipelines using OpenAI, Google GenAI, and LangChain.
• Defined system design standards, data modeling strategy, and API contracts.
• Established scalable background processing with BullMQ and Redis, and implemented observability using Prometheus, Grafana, and structured logging with Pino.
• Owned DevOps setup, CI standards, and code quality enforcement. Coordinated technical decisions, reviewed code, and ensured performance, reliability, scalability and maintainability of the platform.

A mobile application that aggregates odds across multiple sportsbooks and sports, with a primary focus on soccer. Its core feature was a personal betting assistant called Clutch, which interacted with users about matches and players, and supported decision-making through a custom prediction model.

• Primarily responsible for the design and development of Clutch, the AI betting assistant.
• Architected and implemented a multi-agent system where specialized agents handled match retrieval, player analysis, contextual reasoning, and decision support.
• Focused on response reliability, consistency, and maintaining a stable conversational tone aligned with the defined “buddy” persona.
• Integrated LLM workflows using Vercel AI SDK, OpenAI, and LangSmith, and ensured predictable behavior through structured prompting and orchestration logic.
• Contributed to odds aggregation services and implemented integration of Clutch with the React Native mobile application.

A platform that enables users to manage their social media profiles—primarily X (formerly Twitter) and Instagram—through a multi-agent AI system. Users can enable or disable specific agents, such as engagement, research, and content-writing agents, to control automation granularity. The system manages profiles end-to-end, including post creation, comment responses, and performance analytics. It targets social media managers, influencers, and businesses seeking AI-driven profile management.

• Led a team and developed both frontend and backend systems, including the web interface and backend APIs.
• Designed and implemented Python microservices powering the AI agents for content creation, engagement, and analytics.
• Integrated agent orchestration to allow granular control over automation.
• Developed real-time dashboards for post and engagement tracking, and ensured reliable scheduling and execution of AI-driven tasks.
• Optimized system performance for handling multiple profiles and high-frequency interactions, while maintaining consistent user experience and AI behavior.

An email marketing platform that enables users to create rich, interactive email campaigns. Users can include text, images, voting blocks, charts, and other dynamic content. The platform tracks engagement using pixels, collecting data on if, when, and where recipients opened emails, allowing detailed analytics and insights. It targets marketers and businesses seeking data-driven email campaigns.

• Developed both frontend and backend systems to support campaign creation and management.
• Built a rich email editor for customizable templates and ensured efficient handling of large data objects.
• Implemented engagement tracking and analytics pipelines to monitor email interactions.
• Optimized performance and reliability for high-volume campaigns, delivering a robust, full-stack solution.