Work Experience
Data Scientist, Multi-Agent Assistant for Sales Enablement
Duration: 6 months
Summary: Developed a multi-agent assistant for sales enablement, capable of autonomously preparing lead summaries, company/industry reports, and meeting agendas.
Responsibilities:
- Designed and implemented a multi-agent system that used Tavily API for intelligent web search and synthesis;
- Automated generation of sales research reports (for person, company, industry) and meeting agendas based on lead data and found information;
- Built a knowledge base system allowing users to upload files or provide website links for ingestion and querying;
- Developed a RAG interface to let users interact with their documents and websites in natural language.
Technologies: Python, OpenAI, Tavily, Docling, pgvector, SQL, Langchain.
Data Scientist, Legal Assistant using RAG System
Duration: 3 months
Summary:Legal assistant powered by a RAG system for U.S. constitutional and federal law. Enabled users to ask complex legal questions and receive accurate, citation-based responses grounded in the U.S. Constitution, federal law, and New York State regulations.
Responsibilities:
- Prepare dataset by parsing and embedding legal documents using text embeddings model;
- Researched and experimented with different text embedding models from HuggingFace to optimize performance and quality;
- Implemented a process for intelligent query reformulation step using RAG to improve retrieval accuracy based on the user's intent.
Technologies: Python, HuggingFace, Langchain, OpenAI, AWS.
Data Scientist, Virtual Agent for FAQs and Ticket Logging
Duration: 1 year
Summary: Create a virtual agent for FAQs and logging tickets on TechSupport system. Extract names, mails, phone numbers, issues and details during the conversation fill the ticket on system or send self-guide.
Responsibilities:
- Design conversational flow and develop a chatbot for realistic conversation;
- Create a system for managing the connection to the TechSupport system for creating tickets;
- Extract detailed information from conversation and post-process conversational records and text;
- Implement issues classification system for smarter assignment of tickets on the TechSupport side;
- Integrate OpenAI Whisper to improve sound processing.
Technologies: Python, Dialogflow CX, Spacy, NLTK, Twillio, Google Firestore, OpenAI API, GCP.
Data Scientist, DriveED
Duration: 6 months
Summary:Developed a multi-agent generative AI system to automate the creation of lesson plans aligned with U.S. educational standards.
Responsibilities:
- Designed a modular pipeline for generating lesson plans, integrating curriculum standards, success criteria, and textbook-based task generation;
- Implemented multi-agent orchestration for concept-based task design, visual asset generation, and final lesson assembly;
- Tuned agents to adapt lessons by grade level and complexity;
- Integrated RAG to access relevant textbook content and success criteria dynamically;
- Contributed to task formatting and refinement to meet pedagogical goals and improve classroom usability.
Technologies: Python, OpenAI, Langchain.
Data Scientist, Interpretr AIÂ
Duration: 6 months
Summary: A mobile app designed for recording and interpreting users' dreams through a psychoanalytic lens. The bot engages users in interactive discussions to gather insights into their dreams and take into account their life context, aiming to provide meaningful interpretations.
Responsibilities:
- Crafting a natural dialogue flow to mimic friendly and insightful real-time conversations;
- Building a system for collecting and analyzing detailed information on dreams, and providing interpretations rooted in Jungian theory;
- Set Up ElevenLabs with cloning and tuning;
- Upgrade pipeline for real-time Text to Speech;
- Integrate Image Generation pipeline with Flux for visualization of Dreams;
- Image generation with Flux.
Technologies: OpenAI, Flux, Python, ElevenLabs, Grok, Claude, langchain, AWS.
Data Scientist, Tech Cargo System
Duration: 1 year
Summary: The project focused on analyzing and extracting financial information from documents to assess the financial health of companies. Supported multilingual and multi-format document processing.
Responsibilities:
- Leveraged AWS Textract for automated data extraction from financial documents, including balance sheets, income statements, and cash flow statements;
- Designed workflows for financial health analysis using custom formulas and algorithms;
- Integrated OpenAI Visual Language Models for testing advanced data extraction capabilities;
- Built multilingual pipelines using AWS APIs;
- Enhanced support for diverse document types, ensuring scalability and robustness.
Technologies: Python, AWS Textract, AWS Textract Queries, OpenAI API (Vision and Text), AWS Translate, AWS Currency Converter.
Education
- Master of Computer Science
- Bachelor of Computer Science
Upstaffer since April, 2025
