Bence Szabó

Tech Lead & Full-Stack Engineer | 7+ Years Building Scalable Systems | Team Leadership & AI-Enhanced Development Workflows

Table of Contents

About

Senior Software Engineer with 7+ years driving technical innovation in high-growth startups. Proven track record in building scalable microservices architectures, leading distributed teams, and achieving measurable productivity improvements through AI-assisted development. Specialized in backend systems, DevOps automation, and creating maintainable architectures that scale. Led successful migrations of 15+ repositories into unified monorepo architectures and established deployment protocols that enhanced system security and reliability.

Experience

BeamLive

2025.01 - Present: Tech Lead

Key Initiatives:

Culture and Mentorship:

Skills: Leadership · Microservices · NestJS · Redis · TypeScript · AI Agents · MCP · Prometheus

2022.02 - 2024.12: Senior Software Engineer

Key Projects:

Skills: TypeScript · NestJS · Redis · Jest · Microservices · Multi-Tenant Management · Message Queue

2020.08 - 2022.01: Full Stack Software Developer & Product Owner

Skills: NestJS · TypeScript · MongoDB · PostgreSQL · Redis · Angular · React · MQTT

CodeBerry Programming School

2018.10 - 2020.08: Software Developer

Core Development Areas:

Skills: Node.js · MongoDB · GraphQL · API Development · Docker · JavaScript · React.js

2018.06 - 2020.08: Data Analyst

Key Analytical Functions:

Skills: Data Analytics · MongoDB · Google Data Studio · Google Analytics

2018.09 - 2019.03: Code Mentor

Guided and mentored students in coding practices, helping them build practical skills through projects in web development.

Skills: Cascading Style Sheets (CSS) · JavaScript · HTML

2017.06 - 2018.06: Data Scientist Intern and Machine Learning Engineer

Key Contributions:

Skills: Machine Learning · Node.js · MongoDB · Data Analytics · R · Artificial Intelligence (AI)

Freelancing

2018.08 - 2020.08: Full Stack Developer with Data Analytics experience

Skills: Python · Next.js · Node.js · MongoDB · React · JavaScript

DreamJobs

2019.03 - 2019.05: Data Analyst

Skills: Data Analytics · Google Data Studio · Google Analytics

Welcome.Chat

2018.09 - 2019.02: Data Analyst

Skills: Data Analytics · Google Analytics

Dmlab

2018.07: Data Engineer

Skills: Python · Data Analytics · Git

Budapest University of Technology and Economics

2016.09 - 2016.12: Project Management Teaching Assistant

Skills: Project Management

SANXO

2015.06 - 2015.08: Software Developer Intern

Developed a computer vision program utilizing advanced multi-integration camera technology to measure the speed of any moving object.

Skills: OpenCV · Computer Vision

Education

Budapest University of Technology and Economics

2016 - 2018: Master of Science (MSc), Electrical Engineering

2012 - 2015: Bachelor of Science (BSc), Electrical Engineering

Certifications

AWS Certified Solutions Architect – Associate (Course Completed)

UdemyCompleted 2024

Course link

Overview: Completed the full preparation course for the AWS Solutions Architect Associate certification, gaining hands-on understanding of AWS cloud infrastructure, deployment models, and architecture design.
Skills: AWS · Cloud Architecture · EC2 · S3 · RDS · Lambda · CloudFormation · IAM · Networking

Learn Ethical Hacking From Scratch

zSecurityIssued Jan 2021

Credential ID: UC-93247c58-26b9-441c-8e99-aab40fd05cd3

Skills: Ethical Hacking

Go: The Complete Developer's Guide (Golang)

UdemyIssued May 2018

Credential ID: UC-VO6G9G6L

Skills: Golang

Convolutional Neural Networks

DeepLearning.AIIssued Jan 2018

Credential ID: PQR3G7RXKTQZ

Skills: Artificial Intelligence (AI) · Machine Learning · Deep Learning · Computer Vision

Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization

DeepLearning.AIIssued Jan 2018

Credential ID: 8B2LKZHCJETV

Skills: Deep Learning · Machine Learning · Artificial Intelligence (AI)

Neural Networks and Deep Learning

DeepLearning.AIIssued Jan 2018

Credential ID: RTTTM2G3S9YQ

Skills: Machine Learning · Deep Learning · Artificial Intelligence (AI) · Python

Projects

2025.09: Deployment Protocol & Security Hardening

Unified all repositories under BeamLive GitHub organization and established robust deployment practices:

Outcome: Enhanced system security and deployment reliability through standardized processes and comprehensive review procedures.

Skills: DevOps · Security · CI/CD · Release Management · GitHub · Policy Design

2025.08: MCP Server Integration

Deployed Model-Context Protocol servers to enable AI agents to interact with core services:

Outcome: Successfully created a foundation for AI-assisted development workflows and autonomous testing capabilities.

Skills: AI Agents · MCP · API Integration · Automation · Protocol Design

2025.07 - 2025.10: Dynamic Model Loading & Schema Generation

Project Overview:
Developed a runtime schema system for Beam’s dynamic data models, unifying metadata and automating schema management.

Key Contributions:

Outcome: Drastically reduced schema maintenance time and improved consistency across services.

Skills: JSON Schema · Runtime Validation · NestJS · TypeScript · Automation · Database Indexing · Schema Design · Redis · Metadata Management

2025.07 - 2025.09: Beam Live Monorepo & DevOps Ecosystem

Consolidated 15+ repositories into a single Beam Live Core Monorepo:

Outcome: Simplified maintenance, unified development workflow, and dramatically improved developer productivity through consolidated codebase.

Skills: Monorepo · DevOps · Prometheus · AI Agents · NestJS · TypeScript · GitHub Actions · CI/CD

2025.03 - 2025.08: Beam v2.2 Migration & System Stabilization

Project Overview:
Directed a large-scale migration project to enhance scalability, testing, and deployment consistency.

Key Contributions:

Outcome: Successfully migrated all core services to v2.2 with improved stability, test coverage, and development processes.

Skills: Microservices · NestJS · Redis · CI/CD · Testing Automation · Integration Testing · Jest

2025.02 - 2025.10: AI-Assisted Development & Team Management

Project Overview:
Integrated AI tools into daily workflows, enabling automated code review and refactoring.

Key Contributions:

Outcome: Dramatically improved team productivity and code quality through systematic AI integration and effective remote team management.

Skills: Leadership · AI Tooling · Remote Team Management · Process Automation · Mentorship · Code Review

2024.03 - 2024.10: Beam CI/CD and Tests

Project Overview: This project focused on establishing a Continuous Integration and Continuous Deployment (CI/CD) pipeline to streamline development and deployment processes.

Created a proof of concept for the CI/CD flow, set up GitHub workflows and configurations to automate several tasks on every pull request (PR):

Testing Framework

By automating these processes and establishing a robust testing framework, the project significantly improved the efficiency and reliability of the development and deployment workflow.

Skills: Integration Testing · Jest · GitHub · Redis

2022.11 - 2024.10: Live 2.2

Project Overview: This project involved the comprehensive restructuring and enhancement of a software platform to improve performance, scalability, and developer collaboration. The initiative focused on managing custom client logic through nano-services and transitioning to a more efficient data management system.

Technical Enhancements:

Team Collaboration and Management:

Project Leadership and Strategy:

Through these efforts, the project not only enhanced the technical infrastructure of the platform but also improved team dynamics and streamlined development processes, setting a strong foundation for future growth and scalability.

Skills: NestJS · Redis · TypeScript · Multi-Tenant Management · JSON Schema · B2B2C · Microservices

2022.04 - 2024.03: Evolution of Location Service and Geofence Management

The project started with the development of a central location base storage (CLBS) system to manage IoT device location data. Initially, locations were stored as GeoJSON in MongoDB, transmitted via HTTP, and retrieved by clients to display on maps. This setup was inefficient for real-time tracking and lacked advanced geofencing features.

Key Milestones

  1. Introduction of Central Location Service

  2. Refactoring for Real-Time Efficiency

  3. Optimization and Simplification

Key Achievements

Outcome: This evolution transformed the location service into a robust, real-time solution capable of supporting complex IoT ecosystems and advanced geofencing requirements.

Skills: Node.js · NestJS · Redis · Tile38 · Geofencing · MQTT · GeoJSON

2023.11 - 2024.01: Horizontal Scaling

Project Overview: This project aimed to enhance the scalability of services through horizontal scaling, achieved in two main phases with distinct approaches.

First Version - The initial approach involved:

Second Version - The second phase simplified the architecture by:

Skills: Message Queue · MQTT · Load Balancing

2023.04 - 2023.07: Client Code Reduction

Project Overview: This project aimed to standardize the programming language across all services to enhance efficiency and maintainability. Initially, various services were written in plain JavaScript, which led to inconsistent data structures and interfaces.

Transition to TypeScript: By transitioning all services to TypeScript, the project introduced a more robust development environment that allowed for better design of data structures and interfaces, enabling a form of test-driven development. Type definitions were created and shared between the server and client using a common model types npm package, ensuring consistency across different parts of the application.

Migration to React Native: Given budget constraints, which precluded hiring separate developers for JavaScript, Java (Android), and Swift (iOS), the project explored migrating to React Native. This approach aimed to unify the technology stack (server, database, browser, and mobile clients) under TypeScript. The proof of concept for React Native demonstrated the potential to speed up development by sharing code across different platforms and reducing the need for a large team of specialized developers.

Skills: React · npm · JavaScript · TypeScript

2023.01 - 2023.06: Programmable Admin Platform

Project Summary:

Skills: React · Redis · MQTT · TypeScript · NestJS

2022.08 - 2023.02: Nano-Service Engine and Editor

Project Overview: The project focuses on improving the workflow development strategy for Beam, a platform designed to enable cooperation between organizations, users, and IoT devices within a B2B2C model. Beam aims to provide core services through a chat-based application, leveraging business logic to connect various stakeholders. However, there are challenges and critiques around the previous workflow design, primarily concerning its understandability and effectiveness for developers.

Situation: Beam's strategy is not to develop applications itself but rather to offer services to third-party developers and end-user organizations. Currently, Beam is developing an application to test core functionalities and showcase platform capabilities. There is internal debate about the platform's construction, with differing views on whether to use an SDK or API-based approach for third-party development. The applications currently in use are not sufficiently generalized to meet varying customer requirements, necessitating a more adaptable solution.

Challenges and Concerns:

  1. The complexity and limited control of the previous workflow design.
  2. The architecture is difficult for even in-house developers to understand, leading to suboptimal solutions.
  3. The existing development and debugging tools for workflows are seen as less effective compared to traditional IDE toolchains.

Achievements and Future Proposals:

  1. Separation of core backend functions and business logic through workflows.
  2. Development of a web editor and actionflow logic for managing application and organization-specific UI scripts.
  3. Introduction of TypeScript (TS) support in the workflow and actionflow IDE.
  4. Testing and performance improvements for the workflow server, enabling the management of thousands of active workflows.

Skills: Microservices · NestJS · Redis · TypeScript

2021.11 - 2022.11: Live 2.0

Refactoring and Modularization: The project began by separating the responsibilities of each microservice to ensure efficient and scalable development. Responsibilities were divided so that each database was accessed and managed by a single service, removing global database queries. Additionally, the logical layers within each microservice were separated into controller, service, core, and database layers.

Transition to Microservices: Following a comprehensive redesign of the product, the previous monolithic structure was replaced with a microservices architecture. NestJS was chosen for running these microservices, providing a cohesive and scalable structure for backend services. The adoption of a microservices approach allowed for modular and independent deployment of services, enhancing system resilience and flexibility.

Modernization Efforts: In collaboration with a seasoned system architect, the project aimed to modernize the technology stack and optimize service delivery. TypeORM was leveraged for seamless data management and object-relational mapping, while React and TypeScript were used to develop robust client-side applications.

Server-Side and Backend Improvements: The server-side development was powered by NestJS, offering a rich ecosystem for building scalable applications. The backend was orchestrated on the Node.js runtime, known for its non-blocking, event-driven architecture, ensuring the system could handle high-throughput requirements effectively.

Skills: Microservices · TypeORM · NestJS · Node.js · React · PostgreSQL · TypeScript

2021.09 - 2022.09: Scalable Multi-Tenant Service for a B2B2C Company

The project addressed a critical scalability issue in a B2B2C company, which initially created a separate repository for every customer. This approach led to exponential increases in development, maintenance, and deployment times, making it unsustainable.

Problem:

Solution: A multi-tenant architecture was implemented to enable multiple customers to use a single service while maintaining distinct setups. Key steps included:

  1. Initial Multi-Tenant Attempt
  2. Optimized Multi-Tenant Design
  3. Custom Business Logic for Customers

Key Achievements:

Outcome: The project delivered a scalable, efficient multi-tenant architecture, dramatically improving speed and operational efficiency while enabling customized workflows for diverse customer needs.

Skills: NestJS · MQTT · Microservices · Multi-Tenant Management · TypeScript · Redis · PostgreSQL

2021.11 - 2022.06: Central User Database

Project Overview: This project aimed to address the issues of data redundancy and disorganization in the previous system. The original setup suffered from inefficient storage utilization due to multiple databases managed by different services, which were not coherently designed. To resolve these issues, the project focused on restructuring the user data architecture and transitioning from MongoDB to PostgreSQL, and eventually to Redis, for enhanced efficiency.

Challenges and Solutions: The redundancy in user data was introduced during the initial version of the system, where data was dispersed across multiple databases. The main goal was to centralize this user data into a single, centrally managed database to streamline operations and improve performance. PostgreSQL was chosen over other options like MySQL due to its better fit for the project's needs. The transition from a NoSQL (MongoDB) to an SQL (PostgreSQL) database required introducing a new data structure, where documents were split into well-designed tables. This restructuring also created opportunities for new features and enhanced the system's overall efficiency.

Technical Evolution: The project explored moving core business logic into SQL embedded functions to leverage PostgreSQL's capabilities. Additionally, Redis was adopted later on to further improve efficiency. A significant aspect of this transition was the focus on creating a very fast core service to manage the centralized user database. The introduction of a new data architecture and the shift to SQL databases marked a critical step in optimizing the system's performance and ensuring more organized data management.

Skills: TypeORM · NestJS · Node.js · PostgreSQL · SQL · REST APIs

2022.01 - 2022.04: Core Packages

Project Overview: This project aimed to eliminate code duplication and improve scalability by restructuring the product into packages. Initially, the product contained significant code redundancy, with separate repositories created for each customer, which complicated maintenance and hindered scalability.

Refactoring Efforts: The solution involved creating an npm Angular module library, extracting different modules so that main applications could selectively utilize them. This approach was mirrored on the backend with an npm library comprising NestJS modules used by various services connected to Angular applications.

Introduction of Microservices: Following a complete product redesign, the initial repositories and packages were abandoned in favor of microservices with distinct functions. A core npm package was established to consolidate all essential tools used by these microservices. The core package's version number indicated the overall product version, ensuring consistency.

Further Modularization: As the project evolved, the core package was eventually divided into multiple packages. One package retained core functions while others extracted data model interfaces, schemas, and types, facilitating shared interface definitions across both backend and frontend. This modularization ensured better maintainability, scalability, and efficient development processes.

By refactoring the product into well-defined packages and introducing microservices, the project significantly improved the architecture, making it more scalable and easier to maintain.

Skills: React · Angular · NestJS · TypeScript · npm

2020.08 - 2021.03: Admin Platform

Project Overview: The project revolves around the software development and refinement of an administrative platform initially developed at a rapid pace. This swift development necessitated a comprehensive phase of debugging and refinement to ensure system stability and functionality. Alongside this, customer-driven enhancements were identified and integrated into the existing system to meet user needs and expectations.

Development and Enhancements: A significant aspect of the project was the creation of a comprehensive software developers handbook. This handbook was designed to aid new developers in onboarding smoothly and established crucial standards such as code style, commit message, and coding conventions. The introduction of socket communication marked a technical evolution, replacing the traditional HTTP communication method and enhancing system efficiency.

Stakeholder Collaboration: Throughout the project, there was regular and close collaboration with the CEO, who conveyed feature requests from customers. These meetings, held several times a week, ensured that customer feedback was promptly addressed and integrated into the platform. By providing estimated timeframes (ETAs) and continuous feedback on these requests, the development process remained aligned with customer needs and business goals.

Project Outcomes: This structured approach to refining and enhancing the administrative platform not only improved its functionality and user experience but also established a robust framework for future development. The combination of rigorous debugging, systematic documentation, and responsive integration of customer feedback set the foundation for a scalable and efficient administrative platform.

Skills: NestJS · MongoDB · Node.js · Angular · REST APIs · React.js

2019.03 - 2020.08: Google Ads Exporter

The company required a consolidated view of its advertising campaign performance, encompassing both the expenditure and conversion metrics. To facilitate this, developed a Google Script that executes on a scheduled basis—daily, weekly, or monthly—on ads.google.com. This script is designed to aggregate the pertinent data regarding ad campaigns and systematically record it into a designated Google Sheet.

Skills: Google Ads · Data Analytics · Google Apps Script · JavaScript

2019.11 - 2019.12: Landing page for a customer

Developed a modern, responsive landing page for a client, implementing custom design requirements and ensuring cross-browser compatibility.

Outcome: Delivered a fully functional, visually appealing landing page that improved client's online presence and conversion rates.

Skills: React · Node.js · JavaScript · CSS · HTML

2019.09 - 2019.10: Optimizing the website of a customer

Project Overview: This project aimed to improve the efficiency and performance of a customer's site by addressing issues with server-side rendering (SSR) React technology.

Problem Identification: The original implementation of SSR React was misconfigured, causing the entire application to reload with every page navigation. This misconfiguration negated the benefits of single-page application (SPA) and server-side rendering, leading to suboptimal performance.

Solutions and Improvements:

By resolving the initial misconfiguration and leveraging Next.js for static rendering, the project successfully optimized the site's performance, delivering a faster and more efficient user experience.

Skills: Next.js

2019.01 - 2019.07: Landing Page of Programming School

Project Objective: The aim was to transition from a WordPress platform to a proprietary application. A designer had laid the groundwork with the creation of fundamental HTML, CSS, and JavaScript files to establish the visual design, along with minimal server-side logic for page rendering.

Implemented Solution: Successfully migrated all existing functionalities from the legacy site to the new application, which included additional subpages, dynamic content features, localization capabilities, and database integration. Also implemented UTM tracking for marketing analytics. To optimize user experience and ensure effective deployment, conducted A/B testing for certain features, establishing an experimental environment for this purpose. Final refinements were made to enhance site speed, with a focus on improving search engine optimization (SEO). Ultimately, the entire site was converted into a static format to further boost SEO performance during production.

Skills: Node.js · JavaScript · A/B Testing · CSS · HTML

2019.04 - 2019.06: Automatic Email Reports (triggers)

Project Synopsis: The company aimed to engage users by sending automated reminders to encourage product usage, specifically for ongoing learning activities. The strategy involved dispatching tailored reports on a weekly basis, as well as reports reflecting the user's activities from the previous day and from a week prior.

Technical Execution: Developed the server-side logic required to support this engagement initiative, complemented by designing the HTML email templates. The system was programmed to execute on a daily and weekly schedule, intelligently querying the database for users based on their recent activity—whether they were active the day before, in the last week, or precisely seven days ago. Upon identifying the relevant users, the process dynamically compiled personalized emails detailing their recent interactions with the product and dispatched these communications accordingly.

Skills: MongoDB · Node.js · JavaScript

2019.03 - 2019.04: Facebook Ads Exporter

Project Overview: The objective was to replicate the functionality of a Google Ads data exporter, with the focus shifted to extracting data from Facebook Ads. The end goal was to achieve a similar output format in the spreadsheet as that of the Google Ads exporter.

Solution Development: Identified an open-source tool that aligned closely with our requirements. After forking the repository, customized the solution to fit our specific use case, ensuring seamless integration and data export from Facebook Ads into our desired spreadsheet format.

Skills: Data Analytics · Google Apps Script · Facebook Ads · Google Sheets

2019.03 - 2019.04: Google Analytics Dashboard

Project Summary: The company possessed a wealth of data that was underutilized. The role was to organize this data effectively, enabling the company to harness it for informed decision-making processes.

Action Taken: Established a Google Analytics account to systematically gather pertinent data. Additionally, formulated and standardized the usage of UTM tags to ensure consistent and meaningful data collection practices.

Skills: Data Analytics · Data Science

2019.01 - 2019.03: Badges (gamification)

Project Context: The company aimed to enhance user engagement within their application by introducing a gamification feature. They decided to implement a badge system as the initial step towards this goal. The visual assets and designs for the badges were pre-defined and provided.

Technical Contribution: Primary focus was on developing the server-side logic responsible for the criteria-based unlocking of badges. This involved creating algorithms that tracked user achievements and unlocked badges accordingly. Additionally, addressed minor client-side challenges to ensure a seamless integration of the badge system into the user interface.

Skills: API Development · MongoDB · Node.js · JavaScript · REST APIs

2018.11 - 2019.02: Payment Page of Programming School

The company aimed to transition from its existing multi-page payment system to a streamlined one-page application. This initiative was part of an effort to modernize the user interface in line with a new design provided by a professional designer.

Responsible for transferring the functionality of the legacy payment system to the new single-page application.

The project was successfully executed, resulting in a functional and aesthetically pleasing one-page payment application that aligns with the company's vision for a more efficient and user-friendly interface.

Skills: Node.js · JavaScript · REST APIs

2018.07 - 2018.08: Growth Dashboard (AARRR metrics)

Project Contribution: Data Visualization Developer

Project Aim: The growth team required a comprehensive dashboard to visualize user activity and behavioral data, which would serve as a foundation for deriving insights and informing strategic business decisions.

Solution Delivered: Started the development of a sophisticated dashboard that provides a visual representation of user engagement and patterns. Initial task involved scripting a data transformation process that converted raw data from MongoDB into a structured format suitable for chart plotting. This enabled the seamless integration of complex datasets into the dashboard, allowing for intuitive analysis and reporting.

Furthermore, integrated Google Analytics with the dashboard, providing access to key performance indicators such as page views and other relevant metrics. This integration offered a holistic view of user interactions and website performance, thereby enhancing the growth team's ability to make data-driven decisions.

The successful implementation of this dashboard has empowered the company with actionable insights, optimizing its approach to user engagement and growth strategy.

Skills: Data Analytics · Python · Google Analytics

2018.07 - 2018.07: Power Plant Dropout Web Scraper

Project Objective: The company aimed to develop an application capable of monitoring power plant dropouts and notifying users promptly upon their occurrence.

Technical Solution: Responsible for designing and implementing the web scraping component of the project. This involved developing a script that systematically extracts dropout data from targeted websites. The collected data was then formatted and stored in a CSV file, ready for integration with the alert system. This functionality is crucial for the application's ability to provide real-time updates and maintain a high level of service reliability for the users.

Skills: Python · Git

2018.07 - 2018.07: Weather Service Data Fix

Project Challenge: The company faced issues with incomplete weather data collection from various locations, leading to system disruptions due to missing timestamps.

Solution Implemented: Developed a script designed to address the data gaps by filling them with estimated weather data, ensuring the application's stability and preventing crashes. To validate the effectiveness of this solution, also introduced a series of test cases specifically tailored to assess the robustness of the data correction process.

Skills: Python · Git

2018.01 - 2018.02: Automatic Slack Registration

Project Objective: The online programming school sought to expand its Slack-based community by simplifying the registration process. The goal was to eliminate the need for users to navigate a lengthy and unfamiliar form-filling procedure.

Technical Solution: Developed a Node.js application utilizing Puppeteer, a headless browser automation tool, to automate the registration process. The application initially registers using a placeholder email address and subsequently updates the registration to the user's actual email address. As a result, the user receives an email prompting them to set a new password, thereby granting them immediate access to the community's Slack workspace without the hassle of manual form submission.

Skills: Node.js · Git · Slack

2017.06 - 2018.01: Customer Churn Prediction

Project Goal: CodeBerry Programming School required a predictive system to identify users at risk of churning, enabling proactive engagement to retain them.

Solution Developed:

The result was a predictive model that significantly outperformed random selection, correctly identifying users likely to churn with a two-thirds accuracy rate. However, the model's performance was constrained by the limited amount of training data available from the company.

In addition to the analytics work, developed a Node.js application that automated the process of sending retention-focused emails through Intercom to the users identified by the model as at risk of churning.

Outcome: The project delivered a functional predictive system that provided actionable insights, allowing CodeBerry Programming School to engage with users more effectively and reduce churn rates.

Skills: Deep Learning · Artificial Intelligence (AI) · Machine Learning · R · Docker

2017.12 - 2018.01: Rapstar - Generating hip-hop lyrics using neural networks

Project Description: Personal endeavor during university years to create an autonomous rap lyricist. Inspired by a Multi-layer Recurrent Neural Networks (RNN) project discovered on GitHub, the goal was to train a neural network to generate hip-hop lyrics independently.

Technical Execution: Developed a web scraper to collect a substantial corpus of Hungarian hip-hop lyrics from a dedicated lyrics website. After performing the necessary data cleaning to refine the dataset, initiated the training process for the RNN with these texts.

Outcome: The project's end result was a rap-bot that produced text resembling coherent Hungarian lyrics, although the content lacked a meaningful sense. It became evident that achieving the creation of authentic and intelligent lyrics would require a more extensive dataset and prolonged training period.

2016.09 - 2017.06: Smart Retail Product Recognition System

Project Overview: Budapest University of Technology and Economics – Diploma Thesis and Startup Initiative

The project originated as a startup concept and evolved into a comprehensive diploma thesis. The initiative aimed to revolutionize retail operations through computer vision technology, specifically targeting product identification and checkout automation.

Research and Validation:

Technical Implementation:

Analysis and Documentation:

Outcome: Successfully transformed an entrepreneurial concept into a rigorous academic project, demonstrating both technical feasibility and identifying critical challenges for real-world deployment in retail environments.

Skills: TensorFlow · Keras · Google Cloud Vision API · Computer Vision · Machine Learning · Lean Methodology · Team Leadership · Product Research

2015.02 - 2015.05: Teachable Remote Controller

Project Overview: An independent university project focusing on embedded system design and hardware–software integration. The goal was to create a programmable remote controller capable of learning and reproducing infrared signals from other remotes.

Key Contributions:

Outcome: The final device successfully stored and replayed infrared signals from arbitrary remotes, functioning as a fully programmable universal controller.

Skills: Embedded Systems · C Programming · Microcontrollers · Circuit Design · Infrared Communication · Soldering · Hardware Prototyping

Volunteering

Invisible University

2019.01 - 2022.01: Chief Executive Officer

2016.06 - 2019.05: Organizational Development Consultant and Team Lead

2018.09 - 2019.08: Nonviolent Communication Specialist

Offered pro bono training in nonviolent communication (NVC) to young adults, fostering empathetic engagement and conflict resolution.

2018.09 - 2018.11: Gamification Expert

Provided pro bono gamification training to young adults, aiming to engage and motivate through the application of game-design elements in non-game contexts.

2015.09 - 2018.02: Trainer of Trainers

Facilitated comprehensive training programs for the educators at the Invisible University, covering a range of topics including:

2014.02 - 2014.11: Time Management Coach

Conducted time management workshops for undergraduate students, focusing on enhancing their organizational skills and productivity.