Radu Irava

I created a new product line in the digital audio advertising industry. This focused on product-led growth as opposed to direct sales, and its growth contributed to the company’s 9-figure acquisition.

I exponentially grew the fashion eCommerce brand via its mobile apps, contributing to the company’s acquisition. The company became the biggest online fashion player in Central and Eastern Europe.

Oversaw business case, operational plan, pitch deck, investor meetings, and negotiations. The activities resulted in being offered an investment of 600,000 EUR from a strategic investor with experience in the blockchain industry.

▪ Oversaw landing pages structure and content.
▪ Handled mechanism design for beta release.

▪ Defined the product and refined it based on market opportunities.
▪ Provided requirements engineering for mobile apps and complex back-end systems.
▪ Managed communication with software development partners.
▪ Contributed to sales and marketing strategy.

Built a prototype for a web-based visual management tool using 3D spherical camera technology. The software was helping users in the marine shipping and transportation industry with inventory management.

The project represented a sophisticated evolution in personalized language education for a major language learning business (eight figures in users, nine figures in revenue, and four figures employee count). I developed a multi-agent AI architecture using the CrewAI framework to transform the traditional textbook-like learning experience into something more dynamic and conversational.

The system ingeniously integrated multiple data sources, including user onboarding forms, self-assessments, lesson analytics, and performance data, to create a highly personalized chatbot that could handle real-world learning scenarios. The architecture featured specialized agents with distinct roles, including motivational agents and task-specific agents, all orchestrated by a main agent that would parse user queries and delegate different components to appropriate sub-agents. During the initial rollout, the team implemented human-in-the-loop validation to ensure accuracy and reliability, ultimately creating a learning experience that felt more like having a knowledgeable conversation partner rather than working through static educational content.

Built a multi-agent AI system for a major language learning company (100+ million users, $100+ million revenue, and 1,000+ employees) that transforms static textbook lessons into personalized conversational learning experiences using specialized AI agents.

The project emerged from a collaboration between two general managers from major logistics companies and me (9-figure revenue each and 4-figure employee count each). The GMs and their senior executives faced a critical operational bottleneck. Business teams requiring data insights were experiencing weeks-long delays as all requests had to be routed through technical teams for database queries and analysis.

To solve this problem, I developed a sophisticated chatbot solution using LangChain and OpenAI. It could accept natural language queries from business users and automatically translate them into SQL commands, execute them against the company’s ERP database, and return results in an easily digestible format. The roll out started with setting up pilots and and resulted in full implementations at both logistics companies, demonstrating the significant value and impact of democratizing data access across non-technical business functions.

Developed an AI chatbot for two major logistics companies ($100+ million revenue and 1,000+ employees each) that translates natural language queries into SQL commands, eliminating weeks-long delays for data insights and democratizing database access.