Eduard – Python, Vue.js, Django

At first, this was made for language learners with bad eyesight. Later, it evolved into an automated video generator. The product extracts the text from the ebook and groups it into sentences, paragraphs, and chapters. Once the text is processed, it generates audio for each sentence in the designated languages, with configurable (appropriate for learners) speed.

In the case of video generation, Eduard used the knowledge of natural languages (and LLMs) to write precise prompts that will group and generate the image prompts. These image prompts are sent to resolve, and after the images are generated, the program combines the audio and images to generate engaging videos for the language learners.

• Defined the requirements;
• Refined the requirements by rapid prototyping;
• Crafted precise LLM prompts for text extraction, text translation, text grouping, and image generation;
• Implemented the generative workflow with Python, Claude, and Resolve (for image generation).

Designed and implemented an app for language learners who wish to learn vocabulary and grammar like natives do.

• Developed and launched a fully automated, AI-powered language learning app with adaptive flashcards and personalized modules;
• Implemented specialized semantic search with embedding models and pgvector;
• Engineered the prompts;
• Created a multi-agent AI workflow with LangChain (OpenAI + Claude) to analyze user progress and dynamically adjust flashcard difficulty levels;
• Built a scalable backend with FastAPI and Pydantic to manage diverse language content formats (CSV, SQL) and support real-time updates;
• Designed an intuitive user interface;
• Implemented data-driven personalization to tailor vocabulary lists, flashcard suggestions, and practice sessions based on individual learning behaviors;
• Collaborated with AI engineers and language educators to optimize adaptive workflows and enhance content delivery;
• Deployed the app on a scalable cloud platform;
• Created CI/CD pipeline;
• Enhanced retention rates through a multi-level flashcard system that adapts to user proficiency and promotes progressive language development.

Real-time e-commerce price tracking platform

• Architected and implemented an end-to-end automated price monitoring system integrating multiple AI-driven data collection agents and automated decision workflows;
• Designed and built a microservices architecture;
• Implemented intelligent data processing pipelines using Python, integrating with Google Shopping API and custom scrapers;
• Developed automated notification systems and decision support mechanisms to track price changes and market dynamics;
• Created an intuitive Vue front end with advanced filtering, data visualization, and automated reporting;
• Engineered backend services using FastAPI and PostgreSQL to handle complex data processing and automated workflows;
• Implemented automated deployment pipelines and infrastructure management using Docker, GCP, and CI/CD workflows;
• Designed and implemented automated multi-source data collection systems with intelligent proxy rotation and error handling;
• Managed the entire development lifecycle from architecture design to production deployment.

Interactive dashboard for personality data visualization and analysis

• Designed and implemented the product more than half of the year before the deadline by evaluating and elegantly aligning client's future and current product requirements;
• Designed frontend components with a focus on modularity (which, consequently allowed for other desired properties of software engineering);
• Designed and implemented the frontend interface with the user in mind to allow for quick and optimal evaluation of the generated results;
• Solved arising issues in a seamless and efficient manner, making technical trade-offs in such a way as to primarily benefit the client's business and workflow.

The app successfully utilizes GPT and prospect intel (from LinkedIn or .csv notes the user provides) to generate personalized subject lines and emails. It has an application beyond cold outreach, but the User Persona for this project is mostly email marketers who want to create a personalized email campaign - fast!

• Designed and implemented the product single-handedly;
• Devised an user interface and GPT prompt to extract maximum value for minimal effort for the client (and their users);
• Designed and implemented innovative features for personalized bulk email generation;
• Determined the User Persona together with the client;
• Evaluate the competitors, the tech stack, and the user flow of the current product;
• Wrote down improvements for the user flow;
• Prototyped the email-sending feature (Gmail API);
• Added the logic to send and track emails;
• Designed new UI and "user flow";
• Deployed the product and performed all of the necessary quality assurance.

Survival analysis tool

• Collaborated with the founder to gather requirements and provide regular progress updates on project milestones;
• Re-wrote the backend API to allow for further upgrades and performant execution;
• Implemented the GPT prompts used for the purpose of data classification;
• Optimized the GPT prompts to improve classification performance and efficiency;
• Resolved critical bugs and issues to ensure smooth and error-free operation of the application;
• Redesigned and implemented the data model;
• Optimized database queries and improved overall system performance;
• Added new features on both frontend and backend;
• Implemented automated integration tests;
• Designed and implemented cloud infrastructure;
• Employed continuous integration and continuous deployment (CI/CD) pipelines to automate the development and deployment workflow.

This is a cloud Intelligent Document Processing (IDP) platform that helps companies of all sizes minimize the time they spend manually processing documents.

• Designed and developed a new FastAPI email processing microservice;
• Troubleshot and improved an internal migration application, streamlining the data transfer process;
• Deployed service on Kubernetes cluster;
• Implemented an integration test suite on GitLab CI/CD;
• Developed and updated unit tests;
• Utilized the ELK stack to identify and resolve complex bugs, improving overall system stability and performance.

The purpose of the app is to ease customer due diligence during the onboarding process for both companies and customers by keeping customer data collected in one place and shared across companies on a need-to-know basis.

• Developed the back-end of the web application, including designing a middleware layer to connect the front-end to the database SQL function interface;
• Added new API endpoints and updated existing ones to improve the application's functionality;
• Re-designed parts of the database schema and created database migrations to optimize data storage and access;
• Updated and debugged front-end features;
• Implemented an automated testing suite and standardizing seed data used for prototype deployments;
• Wrote a program to generate all necessary SQL interfaces from OpenAPI specs and generated test suites for all >150 API endpoints using Pytest, Pydantic, and Jinja templates.