AiACTIV - DATVIETVAC

* Responsibilities:

- Played a key role in developing ad management platforms like Ad Servers, Demand-Side Platforms (DSPs), and Data Management Platforms (DMPs).

- Worked as an integration developer to optimize advertising strategies and audience targeting with various partners.

- Collected and analyzed user activity data, such as impressions, views, and ad clicks, to provide insights and support decision-making.

- Assisted the data science team in building and deploying segment models based on user behavior and preferences.

* Key Contributions:

Built and launched platforms like Ad Networks, DSPs, and DMPs, making data collection and analysis easier.

Developed advanced techniques to collect detailed user activity data, enabling precise audience profiling and targeting.

Contributed to the creation of segment models that offered valuable insights into user groups and preferences, improving targeted advertising campaigns.

* Technologies Used:

Languages & Tools: Golang, Python, Shell Script

Databases & Systems: Aerospike, Postgres, Redis, Mongo, Timescale, Apache Arrow vs. Parquet, Proto Message

Data & Workflow Management: Apache Airflow, DagU, Argo, Kubernetes (K8S), S3, GCS

Messaging & Monitoring: Kafka, Grafana, Elasticsearch

CI/CD & Advertising Standards: GitLab CI, VAST, Prebid, OpenRTB

Work with: Google Ad Manager, Google Ad Exchange, DV360, Yandex, Pubmatic, The Trade Desk, ADSP

Geniebook

* Responsibilities:

Collaborated with Data Science team to operationalize AI models using BERT model.

Managed input features and model versions, ensuring smooth functioning.

Provided access to end-users, facilitating seamless utilization of AI capabilities.

* Key Contributions:

Orchestrated the integration of AI models into operational pipelines, optimizing performance and efficiency.

Implemented version control processes for model iterations, ensuring accuracy and reproducibility.

Streamlined user access to AI features, enhancing user experience and satisfaction.

* Technologies Used: Python, Golang, GRPC, BERT Model, Jupyter Notebook, Google Kubernative, GCP, Airflow, PubSub Service, BigQuery,Gitlab CI, Bazel, Nginx.

Mobile World Group (TheGioiDiDong)

*Responsibilities:

Lead the development of a chat application for company-wide use, enhancing customer and internal communication and collaboration.

Collaborated closely with the Data Team to gather customer data, leveraging data streaming techniques to analyze messages for sentiment (negative and positive).

Designed and implemented input features manager and model version controller to efficiently manage versions of the BERT model, utilized for customer address identification and food interest prediction.

Managed the output of the model to determine the optimal path from supermarkets to the respective customers.

*Key Contributions:

Led a team in the successful development and deployment of the chat application, streamlining internal communication processes.

Implemented advanced data streaming techniques for real-time analysis of customer messages, enabling proactive responses to customer attitudes.

Developed a version control system for the BERT model, ensuring seamless integration and deployment of updated versions.

Optimized the model output to facilitate efficient delivery of products from supermarkets to customers based on their preferences.

*Technologies Used:

Java, Python, Golang, Rust, Asterisk, GRPC, Websocket, Nat Straming, Nat IO, Kafka, Scylla - Cassandra, PyFlink,PySpark, Bert Model, Kibana, Zabbix, Jenkin, Postgres, Gitlab CI, Nginx, HAProxy, MLFLow, Dremio, Pandas, Numpy,Scikit-learn, Tensorflow, Keras, Rasa.

*Published: Customer sastisfaction with a named entity recognition (ner) store-based management application based on customers to customers-support text messages submitted to "Multidisciplinary Applications of Computer-Mediated Communication"

ANDA LAB - SOONGSIL UNIVERSITY, KOREA

Fault Prediction in Wireless Sensor Networks Testing Process

Wireless Sensor Networks require testing on multiple levels to reduce fault as fast as possible. Several issues can affect the reliability of the wireless sensor networks after deployed such as lack of energy, software malfunction and the harsh environmental conditions. My research proposes a new prediction testing mechanism for Wireless Sensor Networks application base on the automation testing and machine learning system. This work was supported by Institute of Information communications Technology Planning Evaluation (IITP) grant funded by the Korea government (MSIT)

Published: Fault Prediction in Wireless Sensor Networks (CTCon 2019 - Institute of Convergence Technology)

*Technologies Used:

Java, Python, C, C++, Shell, MQTT Protocol, Apache Nifi, Apache Kafka, Apache Flink, OpenWRT, Raspberry Pi, Pandas, Numpy, Scikit-learn.

CINNOLAB – BOOKOKE

* Responsibilities:

Specialized in Smart Home and Smart Storage Solutions, with a focus on Wireless Sensor Networks.

Implemented a rule-based model to manage scenarios in smart homes and storage facilities, optimizing efficiency and functionality.

Orchestrated the collection of sensor metrics from Wireless Devices, performing preprocessing tasks before transmitting data to the cloud system.

Utilized OpenCV and Google Vision to detect towel tags and bottle labels, enhancing data accuracy and reliability.

* Key Contributions:

Developed innovative solutions for Smart Home and Storage environments, enhancing user experience and operational efficiency.

Implemented robust preprocessing techniques to ensure the quality and integrity of sensor data collected from Wireless Devices.

Applying advanced image processing technologies to extract relevant information from towel tags and bottle labels, streamlining inventory management processes.

* Technologies Used: Java, Python, C++, Google Vision, OpenCV, Javascript, Kafka, Rabbit MQ, MQTT Protocol, OpenWRT

FUJINET SYSTEMS JSC

* Responsibilities:

Work with a project focused on face detection and data crawling, aimed at enhancing data collection and analysis capabilities.

Orchestrated the collection of data from various sources, including customer interactions and online sources, to gather contact information such as working domains, email addresses, and phone numbers.

Implemented an innovative solution to map contact information with customer pictures at conferences, facilitating efficient identification and communication.

* Key Contributions:

Developed and implemented a robust face detection system, enabling accurate identification of individuals at conferences and events.

Designed and executed data crawling processes to collect relevant information from diverse sources, enhancing the quality and comprehensiveness of customer data.

Integrated contact information with customer pictures, streamlining the process of identifying and communicating with conference attendees.

* Technologies Used: Java, OpenCV, Selenium, MySQL