Education:-
-
Bachelor of Engineering in Electronics and Telecommunication (2012 - 2016)
- Pune Institute of Computer Technology, University Of Pune.
Experience:-
-
Software Developer at PhonePe(A Walmart Company) (2021 - Present)
- Developing software used by millions of Indian for bill payments and recharge. Tech stack used is Core Java, Spring, Elastic Search, SQL based database.
-
Software Developer at Triple Point Technology (2016 - 2021)
- Developed full commercial software for commodity traders using Core Java, Spring and SQL database.
- Mostly worked on designing complex Valuation algorithms, which help in calculating the profit and loss and helps traders to analyze risk and hedge accordingly.
- Gathered dataset, preprocessed the data and built deep learning models for developing qualitative data analysis used for analyzing large customer feedback surveys.
-
Classroom and Forum Mentor,Project Reviewer at Udacity (2016 - 2019)
- Classroom and forum mentor and project reviewer for Self-Driving Car Nanodegree and Flying Car Nanodegree.
- Beta Tester for Robotics Nanodegree and Flying Car Nanodegree.
- Mentored 300+ international students for Deep Learning, Computer Vision and Robotics.
Programming Skills:-
- Java, Python, C++.
Frameworks:-
- Hazelcast, Spring, Keras, TensorFlow, PyTorch, Sklearn, Numpy, Pandas, OpenCV, Spacy.
Projects:-
-
Recharge and Bill Payment System
- Designed and built scalable and high throughput micro services based system used by millions of Indian users for bill payments and recharge. Used Core Java, Distributed Cache, Elastic Search and SQL based database.
-
Building Valuation Engine
- Designed and built a scalable, high-performance, high- throughput valuation engine for valuation and end of the day process for commodity trading (Oil, Gas and Power). Used in memory distributed cache Hazelcast, Oracle, Core Java and Micro service architecture.
-
Programming a Real Life Self Driving Car
- Wrote ROS nodes to implement core functionality of the autonomous vehicle system, including traffic light detection, control, and waypoint following! Tested the code using a simulator, and then ran it on Carla (Udacity’s Self Driving Car). Repo can be found here.
-
Semantic Segmentation
- Labelled the pixels of a road in images using a Fully Convolutional Network (FCN). TensorFlow and encoder-decoder styled models was used. Encoder used was pretrained VGG model.
-
Behavioral Cloning
- Built and trained a convolutional neural network for end-to-end driving in a simulator, using TensorFlow and Keras. Used optimization techniques such as regularization and dropout to generalize the network for driving on multiple tracks. Many data augmentation techniques were used. Repo can be found here.
-
Path Planning
- Built a path planner that navigates a vehicle through traffic on a highway. Concepts used-environmental prediction, behavioral planning, and trajectory generation - to build the planner. Repo can be found here.
-
Traffic Sign Classification
- Built and trained a deep neural network to classify traffic signs, using TensorFlow. Experimented with different network architectures. Performed image pre-processing and validation to guard against over fitting. German traffic sign dataset was used for training. Repo can be found here.
-
Lane Lines Detection
- Detected highway lane lines on a video stream. Used OpenCV image analysis techniques to identify lines, including Hough Transforms, Sobel Filter and Canny edge detection.Repo can be found here.
-
Vehicle Detection
- Detected cars on a video stream. Used HOG as feature descriptor. Trained the network using Linear SVM. Sliding window approach used for traversing image. Dataset provided by CrowdAI and Autti. Repo can be found here.
-
Agri-bot
- A robot designed for farming purposes. Two modes for controlling: manual and autonomous. Used PS3 for manual mode. For autonomous mode used a camera for detecting the color and then use that color for pet following. All the sensor data was pushed to the cloud for monitoring purposes.
-
32 bit RISC Processor
- Designed the whole processor core using VHDL and simulated the results on Xilinx IDE.
-
Trek-Mate
- Intel Edison used as the embedded part. The system was mounted on the bike. The system has GPS and various sensors like rotary encoders, alcohol sensors, accelerometer, etc. DC motor is used for locking mechanism. Accelerometer is used to detect the jerks. GPS is used to track the bike.All of the data is available on an Android app. Used ThingSpeak and PubNub cloud for IoT.