Saurabh Powar
AI Enthusiast & Passionate Full Stack Developer 🚀
I am a final year undergraduate student at Veermata Jijabai Technological Institute, majoring in Electronics Engineering.
Currently I am working as a Software Developer Intern at GroundUp, India.
My Interests are:
- Machine Learning
- Deep Learning
- Software Development
- Cross-platform Application Development
- Open Source
If you are inquisitive about DL you can read my blogs at hashnode.com/@sp-lit5
Experience
Milvus @ OSPP Summer 2022
Open Source Developer
June 2022 - Present- Developing a Face Recognition Pipeline using Milvus Database, Image Processing Algorithms and Deep learning.
- The procedure includes Face detection using MTCNN, storing the data of Face embeddings on Milvus Database and Face classification.
GroundUp, Inc.
Software Developer Intern
Feb 2021 - Present- Actively developing new features inside the React Native based GroundUp Mobile Applications, and the Admin Panel web application for monitoring the entire ecosystem.
- Fixed issues and bugs in the present builds, optimized workflow of the applications.
- Written Unit tests for testing different components inside the application using React native oriented frameworks like Jest and Enzyme.
iNeuron.ai, Inc.
AI Intern
Dec 2021 - Jan - 2022- Developed a solution to a problem statement Automated Invoice Processing with OCR and Deep Learning.
- Worked on Digitizing an Invoice which includes extracting pertinent information/data from scanned or PDF invoices and transforming it into a machine readable format that is both editable and searchable.
ROS @ OSPP Summer 2021
Open Source Developer
July 2021 - Sept 2021- The Institute of Software Chinese Academy of Sciences and OpenEuler selected me for the coveted Open Source Promotion Plan initiative.
- Developed safety watchdog node in C++ to ensure the mobile robot is acting properly and not about to collide with an obstacle based on lidar, RGBD data.
AIRPIX, Inc.
AI, ML, and Edge Computing Intern
May 2021 - Aug 2021- Under the problem statement, lead the work of detecting lane markers and tracking their consistency using Computer vision and deep learning techniques.
- Trained YOLOv5 models for Road assets objects detection, achieved a reasonable mAP of around 87%.