Software Engineer | Web Developer | Deep Learning Enthusiast

Skills

Programming

Python C/C++ PHP JavaScript

Deep Learning | AI

TensorFlow & Keras PyTorch OpenCV LangChain

Web Development

HTML CSS Bootstrap MySQL Flask Django Streamlit

Experience

Software Engineer, Winmate Jan 2024 - Present
  • Deep learning-based automated optical inspection (AOI) algorithm and system design
  • Integration of AI accelerators into existing company products
  • In-house facial recognition system development
  • Development of LLMs Application
Software Engineer Intern, SparkAmplify Jul 2023 - Oct 2023
  • Study and practical application of LangChain's knowledge and architecture
  • Vector database research and implement
  • Chainlit user interface implement
  • Research and application of LLMs
  • Online real-time chatbot development and testing
Research Assistant, MIPL Sep 2019 - Jun 2020
  • Planned ML solution components for new problems
  • Designed data and annotation requirements
  • Developed training and evaluation code for deep learning CV models
Backend Engineer Intern, Cloud Programming Online Jun 2018 - Feb 2019
  • Designed and built a web content management system
  • Implemented REST APIs for mobile application

Education

Projects

  • Langchain Gemini Chatdocs
    • This is a Language Model (LLM) application built with LangChain + Streamlit, utilizing Google’s powerful language model, Gemini Pro. The application allows users to upload multiple PDFs or web content and generate responses to their queries based on the uploaded content.
  • Stroke- Prediction
    • This project integrates a 3D-CNN model for CT image classification, leveraging YOLOv5 to detect vital organ regions within the images. Uniformizing techniques are applied to standardize input image sizes. Using 3D-Grad-CAM for visualizing model attention on critical areas within the CT scans.
  • Meter Reader
    • Reading single pointer analogue meter using object detection and image processing algorithms, including Hough transform, Edge detection, Polar coordinates mapping, etc.
  • Facial Expression Recognition & pose estimation
    • Based on the principle of CNN, we built an application to classify the expressions on face images into various categories and estimate the pose of a person from a video.
  • Weshare
    • A social media site built with Django and Python, which is similar to Instagram, allows users to edit and upload photos, as with other social networking platforms, users can like, comment on and bookmark others’ posts.