Portfolio

Here are some highlights of my work in AI product management.

Semantic Search Across Multiple Document Repositories

Global Service Provider

Developed secure internal AI-powered semantic search of confidrential data records stored in multiple document repositories on company SharePoint and Confluence.

Key Achievements:

Automated time consuming manual search for information across nultiple document repositories on company Confluence and SharePoint
Improved operational efficiency by 60%
Reduced onboarding time for new employees

Technologies & Methodologies:

Indexed Vector Databases, Natural Language Processing (NLP)
Large language Models (LLMs), Retrieval-Augmented Generation (RAG)
AI Agents, API integration
Analytics, Document Repositories

Multi-Agent Automation Tool for Customer Support

Large European Mobile Operator

Led development of an enterprise-grade conversational AI platform that enabled businesses to automate customer support interactions while maintaining high quality and personalized experiences.

Key Achievements:

Customer support automation tool serves 9M+ mobile and broadband customers
Reduced average time for production issue resolution by 45%
Reduced a churn rate for customers by improving customer support experience
Developed multilingual NLP capabilities supporting 2 languages

Technologies & Methodologies:

NLP, LLMs, RAG
Agile/Scrum, Design Thinking, User Research
Analytics, Performance Monitoring

Generation of Video Highlights Clips for Sport Events

International Sport Club

Led creation of AI chatbot. Application used AI and video editing tools for automatic creation of sport event highlights video or VOD trailers from full length videos of sport events.

Key Achievements:

Increased productivitity of video clips generation for soccer games by 70%
AI-driven highlight editorial workflow is cheaper than human-driven%

Technologies & Methodologies:

Llamaindex framework, RAG, LLM, OpenAI's Whisper
Snowflake Data storage
AWS, CI/CD Pipeline

AI Agent for Workflow Automation

Telecommunications Company

Developed an AI agent to optimize quality assurab=nce workflows. Test cases creation based on requirements stored in Jira and on Confluence. AI Agent can be re-used for multiple projects.

Key Achievements:

Achieved 95% requirements coverage by test cases
ROI realization in about 3 months by reduced manual test design effort and shorter test cycles
Reported productivity gains for test creation and testing by 50%

Technologies & Methodologies:

LLM APIs: Anthropic Claude API, OpenAI API
Orchestration: LangChain, LlamaIndex for complex workflows
Jira Libraries: atlassian-python-api, jira-python
Deployment: AWS Lambda, Google Cloud Functions, Docker containers

Generative AI-powered Content Creation

US Advertisement Company

Led product strategy and development for an advertisement company re-branding and creation of company new website. AI-powered content creation helped marketers and content creators generate high-quality, new brand-aligned content at scale.

Key Achievements:

Launched new attractive website ahead of schedule
Secured partnerships with 6,000 major marketing agencies in 2025
Reduced graphical content creation time by 75%

Technologies & Methodologies:

Large Language Models, Diffusion Models, Prompt Engineering
Content Moderation, Bias Detection
UI/UX Design, Figma
Amazon AWS Cloud, PostgreSQL, Domo BI & Analytics

AI-Powered Personalized Recommendation Engine for Video Content

Video Production Company

Managed development of recommendation engine using a novel approach that combines Retrieval-Augmented Generation (RAG) with multi-agent systems to create more personalized, context-aware suggestions. This method, called Agentic RAG (ARAG), significantly outperforms traditional recommendation systems by using specialized AI agents that work together to understand user preferences and match them with relevant items.

Key Achievements:

Created personalized customer experience for video recommendations
ARAG demonstrated 42.1% accuracy gains across different product categories in NDCG@5 (a ranking quality metric)
Anpnysmous Data Aggregation and Analysis
Enhanced Security controls
Maintained 99.9% system reliability with strict privacy compliance

Technologies & Methodologies:

LLMs, Multiple Agents, Gemini
ARAG, Google Cloud
Vector Databases

Let's Create Something Amazing

Interested in learning more about my work or exploring collaboration opportunities? I'm always excited to discuss new AI product challenges and innovations.

Back to Home