Software Engineer (AI/ML)
Audax Labs
Noida
06.2022 - Current
Project 1 - "OCR Engine release 1.0 & 2.0"
- 1.0
- OCR engine was built for Hitachi Vintara
- Kafka is used to send messages.
- Kafka batch processing implemented.
- Additional enhancements include multi-language support and text translation.
- Parallel processing & multithreading of documents implemented.
- Millions of documents processed using our OCR engine.
- License verification using system id and date, i.e it will only work on system for which license was created.
- 2.0
- GUI and Admin UI support, from which we can manage OCR license, user management, log analysis, provision and deprovisioning of node.
Project 2 - "Reverse Image Search"
- A reverse image search was built for Marvel Studio to search their database .
- It takes a image as input and returns all related images.
- Xceptionet was used for training. Reverse image search is a technology that takes an image file as input and returns related results.
- Dockerized application for better scalability.
Project 3 - "Traffic Monitoring & Accident Detection"
- This solution will enable state DOT/City to deliver geospatial visualizations of road network performance and demand, travel time delay, non-recurring congestion events. Record, analysis, and dispatch crew for pavement repair, and other hazardous conditions.
- Ability to upload a video. Classify vehicles based on vehicle type (Car, Pickup Truck, Bus and Motorbike ) Count vehicles for each vehicle type
- Identify if there is any incident/ accident and notify the user (configurable email address)
- User interface will allow users to visualize the camera locations on the map (if the video has geo information)
- User can select any camera on the map to get the further details about the location.
- The UI will allow users to interface with the video and review the identified features.
- The system will store inference data to Azure based on predefined frequency.
