avatar

Phan Tien Dat

@dnh30701_17937
5.0
Developer
Python C++ SQL Latex
Ha Noi, Viet Nam Tham gia: 05/02/2025 2:43:04
AI/ML Engineer seeking challenging opportunities in the field of Artificial Intelligence and Machine Learning.
Liên hệ

Quá Trình Học Tập

Hanoi University of Science & Technology
Bằng cấp: Bachelor of Applied Mathematics & Informatics
(Gpa: )

Kinh Nghiệm Làm Việc

GetflyChatbot ( - 6/2024-Present)
- Developed and optimized chatbots like Getfly, LPBank, VPBank using FastAPI, PostgreSQL, Gemini, and Streamlit.
- Built FastAPI endpoints to manage interactions and ensure fast, accurate responses, while integrating data embeddings with BGE.
- Worked across development, staging, and production environments, ensuring smooth deployments and system stability.
- Managed all chatbots under a single codebase to streamline maintenance and improve system scalability.
- Built services for the Front-End (FE) to interact with, as well as developed systems for synchronizing customer data across multiple platforms.
AI Engineer
LPBank, VPBankChatbot ( - )
- Scraped data from LPBank & VPBank websites using BeautifulSoup, and processed internal PDF documents for structured storage and retrieval.
- Designed chatbot to handle both RAG-based answers and flow-based interactions, enabling flexible responses based on context.
- Created Dockerfile for containerization and collaborated with DevOps to deploy the chatbot via CI/CD pipeline.
NL2SQLChatbot ( - )
- Utilized Milvus for vector-based search and SQLite for structured data storage.
- Integrated advanced LLMs to translate user queries into accurate SQL, enabling seamless database interactions.
- Implemented Retrieval-Augmented Generation for improved data retrieval and response accuracy.
- Applied SQL normalization and reranking with cross-encoder models to enhance response precision.
- Built FastAPI endpoints for query handling, data retrieval, and response generation.
Speech Processing EnginnerBox Studio (10/2023 - 5/2024)
- Developed and trained Text-to-Speech (TTS) and Voice Conversion (RVC) models.
- Processed voice data to create diverse datasets, improving the naturalness and accuracy of speech synthesis.
- Created generative AI content using tools like MJ, integrating TTS and RVC outputs for media campaigns.
- Managed social media platforms (YouTube, Facebook, TikTok), optimizing reach and engagement.
- Built Playwright-based automation tools to streamline tasks such as content posting, scheduling, and data extraction across platforms (YouTube, Facebook, TikTok).

0 Review