Teodor – Python, OpenAI, LLM

It's a virtual/agentic employee and organizational intelligence platform.

• Led the end-to-end architecture and delivery of the Knowledge Graph Agent, a company-wide LLM-driven platform for organizational risk visibility.
• Owned the product through company-wide release and guided its technical direction.
• Built the ingestion pipeline for Slack, GitHub, Notion, Rippling, and Calendar data into Apache AGE.
• Developed LLM-assisted analysis for risk assessment and executive reporting.
• Established continuous deployment workflows, enabling daily production releases for AI services.

It's an AI-powered platform for hedge fund analytics and portfolio management that enables clients to make informed investment decisions.

• Led development of a data processing platform while mentoring mid-level developers.
• Designed and deployed production RAG pipelines for feature extraction and portfolio analysis in finance systems.
• Built RAG AI agents with LangChain and OpenAI LLMs for advanced hedge fund and portfolio data interaction.
• Architected analytics pipelines using Pandas, Airflow, Databricks, and Spark.

The projects are internal initiatives to streamline the auditing process and improve efficiency across all company member firms by providing auditors with innovative solutions powered by automation, OCR, and AI.

• Led development of an automated financial auditing platform that saved thousands of auditor person-hours.
• Engineered OCR and rule-based document extractors with a validation rule engine for audit compliance.
• Directed a fraud detection system for accounting ledgers using pattern-matching algorithms.
• Automated fraud detection workflows to streamline the overall auditing process.
• Built a full-stack web app unifying auditing and fraud systems with a scalable back-end and an intuitive front-end.
• Integrated real-time data processing and visualization for actionable audit insights.

It's a retail market intelligence platform that helps both businesses and end users make informed decisions about the pricing and availability of retail products (electronics and consumer goods).

• Engineered an internal data warehouse and ETL pipelines tailored for the retail industry.
• Led distributed data normalization and validation pipelines optimized for speed and accuracy.
• Deployed the solution using Docker and Swarm for scalability and seamless updates.

It's a large-scale media monitoring system that enables accurate analysis of a large number of traditional and social media sources.

• Led the company's data collection team and mentored four engineers.
• Scaled distributed crawling architecture from <1K to 100K+ websites hourly, driving media monitoring sales growth.
• Built datasets for language models and NLP applications.
• Contributed to BI platform, CRMs, and data visualization dashboards.