# Self-Driving Car / Automated Navigation of car in an official Ubisoft game using Artificial Intelligence, Reinforcement Learning and Computer Vision

Designed and Implemented self-driving car in an official game in Ubisoft. this project involved complete design & functionality analysis and at the same time assuring its quality using automation testing with AI self-driving agent trained on game environment and perform desired task without manual intervention.

Technologies/Tools used: Python, Evolution Strategy, Reinforcement Learning, Computer Vision | OpenCV, NumPy, TensorFlow | Keras, ELK Stack

Role and Responsibilities:

• Trained neural network using cnn to detect road using image segmentation techniques.
• Trained and generated a segmentation model for prediction of road detection using deep neural

network and achieved 97?curacy on hold-out test data set.

• Implemented image augmentation using OpenCV to increase dataset (46K images)

• Build a model using OpenCV and deep neural networks to help agents identify turns and its distance from game minimap using 5K dataset.
• Implemented and trained agent using multi-agent reinforcement learning for on-road navigation and traverse between multiple points in the map without any manual intervention.

• Implemented evolution strategy, deep learning framework and OpenCV computer vision

using python to generated model to train car to traverse path in openai gym environment.

• Achieved full performance in training environment, as well as previously unseen environment,

through intricate data selection/augmentation strategy and neural network tuning.

• Implemented feature detection for real time object detection and avoid collision and successfully

did open world navigation and drive between two points in the map.

• Out of Box transfer learning of solution to different games.

• Build Reporting Tool to do analysis of data using Elasticsearch, Logstash and Kibana

# Big Data Analysis and Data Science Solution for Jio Mobile Apps

Involved in innovative mobile app product development and integration of mobile apps and conducting basic data & competitive analysis. Blends machine learning skills with great domain knowledge to drive strategy and application of machine learning techniques to real-world problems.

Environment: Big Data Analytics, HDFS, Apache Spark, Machine-Learning, Python, Scala, Hive, PHP, MYSQL, Cassandra, JSON, XML, Android, iPhone.

Role and Responsibilities:

• Involved in implementation of high volume transactional mobile services used by 10M+ customers using Spark and HDFS
• Designed and implemented Spark Jobs.
• Analysis of users’ events and application and content downloaded and time spent details
• Recognizing the themes and topics of the articles and Auto-summarizing text using K-Means and K-Nearest Neighbors.
• Naïve Bayes on application Reviews using Bernoulli
• Sentiment Analysis of tweet
• AD detection using SVM

Ecommerce Chatbot

Chatbot implementation using AI and deep learning that can answer frequently asked question and help customers with details about the product.

Environment: Python, Deep Learning, TensorFlow, NLTK

Role and Responsibilities:

• Trained a chatbot model that respond to customers question and give appropriate response.
• Build Market tracker to process customer request details like product name, product type and product price and provide with the response.