Work on automation testing for network-based test cases, focusing on protocols like Zigbee, Open Thread, Bluetooth, and WiFi. This involved writing scripts to run commands on various device types. Migrate Java test cases to Python-based test cases, with the task starting in January 2024. Perform maintenance work on older codes, which involved fixing scripts with errors from test cases between 2019 and 2022. Conduct maintenance work on servers, addressing issues with executors as needed. These executors are embedded computers located in Budapest and Boston. Investigate test results by checking error logs to determine if issues are related to Software Quality Assurance or development, and create Jira tickets to assign to the appropriate personnel. Manage the team by dividing individual tasks, conducting daily meetings with the client team in Budapest, supporting and training individual members, delivering SQA reports, and handling project documentation. Conduct triages, which involve gathering and running all test scripts from teams in Budapest, Hyderabad, Boston, and Hanoi, assigning the analysis and error sorting to my team, and forwarding results or Jira tickets to the rightful owners.

Developed an Android application tailored for a custom Android device. Implemented object detection technology to efficiently track and record student attendance. Collaborated with educational institutions to ensure the app met school-specific requirements.

Developed an Android app that uses fingerprint authentication to authorize users for pension fund withdrawals. Ensured the application securely manages user authentication and withdrawal processes. Collaborated with the design and backend teams to integrate the app with existing systems.

Used a pretrained Handwritten Text Recognition (HTR) model to recognize handwriting in English exam papers to find spelling mistakes and summarize paragraphs using Skim-lit TensorFlow. Trained a custom model for Bangla Handwriting Recognition for future projects aimed at educating local students.

Engaged with customers from the USA either through direct meetings or emails to showcase the company's current products and services. Detailed customer requirements and communicated them to the developers.

Prototyped a small robot by integrating Jetson Nano, Intel Camera, ultrasonic sensors, lidar, camera motion motor, Arduino Uno Mega, batteries, and robot locomotion motors. Implemented an object detection function on the security robot using the YOLOv3 model.

Created a simulation for a self-driving truck using the Unity engine for the virtualization of a construction site based on LIDAR scanning of a specific map in Japan. This was primarily for testing purposes before real-life implementation. Implemented object detection on the simulated truck. Utilized the YOLOv3 model for object detection and integrated it with the truck's sensors and locomotion systems using the Robot Operating System (ROS).

Implemented the object detection feature on a construction site risk inspection robot using Single-Shot-Detection (SSD) MobileNetV2. Prototyped a small robot by integrating various components including Jetson Nano, Intel Camera, camera motion motor, Arduino Uno Mega, batteries, and robot locomotion motors.

Implemented a Natural Language Processing Model on a Japanese government website, handling a large dataset using the RASA Framework. Configured the NLP model based on English language data and integrated the Google Translation API for translation to Japanese. Developed the website's backend using Django.

Made a line detection model to check red lines on construction yards or buildings for quality purposes. The model was implemented on a small device that included integration of Raspberry Pi 4 and a camera. Developed an object detection model using YOLOv3. Automated the Linux OS on Raspberry Pi by scripting with bash. Modified garp2 to auto-connect all cameras through Wi-Fi upon device startup. Configured the system to send feedback and summary reports to a server at specified times.

Used Py-Tesseract to extract information from CVs and business cards for recommendation systems.