Hello!
Unay Shah
Software Developer at Wizerr AI.
MS Robotics at University of California San Diego.
Interests: Computer Vision, Deep Learning, Software Development.
Experience
Wizerr Logo
Wizerr AI
Jan 2024 - Present
San Jose, CA, USA
Software Developer
A young fast paced startup building the AI Co-pilot for the electronics industry using the latest in ML technologies.
Ria Money Transfer Logo
Ria Money Transfer
Jun 2023 - Sep 2023
Buena Park, CA, USA
Software Intern
  • Developing a Python application with Flask to efficiently identify and resolve server errors.
  • Leveraging Opsgenie to access failure logs, the program filters issues and runs predefined shell scripts to autonomously fix server issues. Deployed the application on AWS ECS and using GitHub self-hosted runners.
  • Enabling quicker response to server issues and replacing manual intervention by Network Operations team.
  • Using GitHub actions for CI/CD to create and deploy Docker images and run scripts on runners.
  • Built an architecture to allow flexible addition of error descriptions and commands to fix issues.
  • UCSD
    University of California, San Diego
    Mar 2023 - Sep 2023
    San Diego, CA, USA
    Graduate Researcher
    Conducting research under Prof. Cheolhong An to obtain 3D cell models from Phase Contrast microscopy images.
    OmniSync Incorporated Logo
    OmniSync Incorporated
    Jan 2023 - Mar 2023
    San Diego, CA, USA
    Software Intern
  • Built a Python pipeline to transfer Segment Data logs from Amazon S3 Buckets to an AWS Postgres Database.
  • Parsed and organized the data into SQL tables, enabling querying and analysis of metrics for 4 teams.
  • Successfully processed and transferred over 300K logs, ensuring efficient and accurate data transfer.
  • Deutsche Bank Logo
    Deutsche Bank India
    Aug 2020 - Jun 2022
    Pune, MH, India
    Software Developer
  • Developed an image data extraction application for 50+ document types (US tax forms, loan contracts etc.), serving 10+ teams and processing 1000+ forms daily using Java, Python, OpenCV and Tesseract OCR.
  • Independently built a Java backend to create form templates and identify form types using image similarity.
  • Utilized OpenCV with YOLO and MRCNN in Python for extracting paragraphs, boxed letters, checkboxes, barcodes, signatures, and bordered tables.
  • Implemented signature forgery detection model using Siamese Networks and SIFT Feature Matching, validating signatures on 1000+ documents every hour with 84% accuracy.
  • Led engagement with 6 client teams, demonstrating the application's capabilities, gathering valuable feedback, and spearheading customized feature implementations to meet their specific requirements.
  • Oversaw and contributed to development of a ReactJS web application with Java Springboot and Python Flask APIs for data extraction and file management services, deployed and maintained on OpenShift Pods and GCP.
  • Mentored 3 interns and 4 junior employees, overseeing projects involving spaCy NLP, GCP Document AI, handwritten text OCR, and SciTSR table extraction.
  • Samsung Logo
    Samsung Research Institute
    Jan 2020 - Jun 2020
    Bengaluru, KA, India
    Software Intern
  • Optimized the retrieval, storage and processing of Samsung Payments log files for error monitoring in AWS S3 Resilient Distibuted Dataset (RDD).
  • Created filters for Spark to process log files. Designed a partitioning schema to improve log file retrieval.
  • Built a dashboard in Scala to monitor service health and error frequency.
  • Deutsche Bank Logo
    Deutsche Bank India
    May 2019 - Jul 2019
    Pune, MH, India
    Software Intern
  • Developed a template creation algorithm to help templatize forms and identify similar forms for data extraction
  • ZConnect Logo
    ZConnect - Hyperlocal
    Aug 2017 - May 2019
    Goa, India
    Graphic Designer
  • Designed logos, T-shirts and in-app elements with Adobe Illustrator.
  • Clicked and used stock photos to create marketing material with Photoshop.
  • LeNest Logo
    LeNest Hospital
    May 2018 - Jun 2018
    Mumbai, MH, India
    Intern
  • Used Google Apps Script to build an appointment reminder system to send e-mails and SMS to patients.
  • Assisted in YouTube content creation and digital marketing by designing thumbnails and clicking photos.
  • Skills
    Python
    Python
    Java
    Java
    C++
    C++
    JavaScript
    JavaScript
    PyTorch
    PyTorch
    TensorFlow
    TensorFlow
    OpenCV
    OpenCV
    NumPy
    NumPy
    React
    React
    Spring
    Spring
    Flask
    Flask
    Android
    Android
    Git
    Git
    GitHub
    GitHub
    MySQL
    MySQL
    TailwindCSS
    TailwindCSS
    Projects
    Fast Class-Based Neural Style Transfer
    Fast Class-Based Neural Style Transfer
    Implemented a real-time class based style transfer using Fast Style Transfer and Fast SCNN-based semantic segmentation. Reduced model size to 1.64MB and improved runtime speed to 25 fps on Tesla T4 GPU.
    Basketball Game Tracking with Faster R-CNN and Team Segmentation
    Basketball Game Tracking with Faster R-CNN and Team Segmentation
    Used Faster R-CNN to track elements in basketball games, conducting ablation studies on the model and datasets. Performed and compared various team segmentations methods.
    Visual Inertial SLAM
    Visual Inertial SLAM
    Used IMU and image feature data to perform SLAM on a car using Extended Kalman Filter to optimize position.
    Particle Filter SLAM
    Particle Filter SLAM
    Performed SLAM of a differential drive robot in a room equipped with Encoder wheels, IMU and LiDAR data, using particle filtering to improve results. Used Kinect data for texture mapping to project images captured by the robot on a map of the floor traversed by the robot.
    Panorama Creation from IMU data
    Stabilized linear velocity data from Inertial Measurement Unit (IMU) and verified results with Vicon data. Used the resulting measurements to track camera orientation and build panorama images with projection mapping.
    Determination of Toxicity of Metal Oxide Nanoparticles Using Regression, Supervised by Dr. Sukanta Mondal
    Built linear regression models with Sklearn in Python to identify properties responsible for toxicity of metal oxides. Studied correlation between properties to reduce dimensions and predict a numerical value of toxicity.
    Countdown Timer Android App
    Countdown Timer Android App
    Built an Android app in Java to run consecutive custom interval timers useful for workouts.
    Education
    UCSD
    University of California San Diego
    MS Robotics
    Sep 2022 - Jun 2024
    BITS
    Birla Institute Of Technology And Science
    BE Electrical & Electronics
    Aug 2016 - Jun 2020
    Hobbies
    Photography
    I have been clicking since 2014 and currently use a Nikon D5600. I am always up for exploring and clicking new places. Here are a few of my clicks.
    Ultimate Frisbee
    I started playing Ultimate with BITS UFC in 2016. I served as the team coordinator in 2018-19, and organized India West Zone Sectionals tournament in 2018.
    Contact Me
    Send me an Email:
    GitHubLinkedInInstagramFacebook