Avani Gupta

I am good in solving problems .I am into Machine Learning, Artificial Intelligence , Reinforcement learning, NLP ,Computer Vision , app development, web development.I know Python , C , Java, Javascript, HTML/CSS, Designing, Java-Netbeans.
The idea of mimicking the brain always fascinates me.My aim is Solving real life problems for the benefit of people.I was curious to know how childrean learn,how human brain functions so well from my childhood. Now seeing AI and reinforcement learning is like a dream come true.Though we in 2019 are far away from making an exact replica of humans in terms of funtinality , adaptability we have started heading towards it.Sophia -A humanoid is already made.I studied the way it talks and algos and methods behind it,which shows we need to go a long way to make human like creatures. Was fascinated while using Reinforcement learning to make Autonomus car, seeing AI play in the environment was awesome. Developed AI playing games like Doom and Breakout. Made Chatbot using deep learning and NLP. Have also worked on object detection and image processing.
I am creative, Studious, self-disciplined , puntual, responsible and deligent. I can work on a problem with full enthusiasm continously without getting tired. Working on Re-inforcement learning and Deep-Q- Learning never feels a work to me.I like to play with Algorithms to optimize a result.
Some of my challenging proejcts are:

• Self Driving Car
  Nov 2018 - Dec 2018
  https://github.com/avani17101/Self-Driving-car.git
  Used Deep Q-Learning . The states of the environment are encoded by a vector which is passed as input into the Neural Network.The best action to play is chosen by highest Q-value or Softmax function. The part to optimise the values like temperature , reward, punishment was quite challenging. Had to do a lot of experimentation on the model for coming up with optimised values. The best part was to see my Car run autonomously in simulations.

*Chatbot
Nov 2018 - Dec 2018
https://github.com/avani17101/CHatbot.git
Chatbot made using Deep Natural Language Processing. Trained Seq2Seq model. Cornell movie-dialogs corpus Dataset used for Training. Again finding the optimal values of parameters was the most challenging and best experience was talking to the chatbot I created! Learnt a lot of optimization techniques and to keep patience when things don’t work the way we want it.

*Artificial Intellingence Playing Doom Game
Dec 2018 - Dec 2018
https://github.com/avani17101/AI-playing-Doom-Game.git
Artificial Intelligence Playing Doom Game: Developed AI which learnt on its own the tactics to play and win in Doom Game. Built Using Deep Convolutional Q-Learning model with Eligibility Trace(n-Step Q-Learning)

*Drinking Liquid purity detector
Sep 2017 - Dec 2017
Used to predict purity of drinkable liquids like water and milk. Consists of a parallel plate Capacitor. Liquid sample is poured in between the plates ,capacitance is measured and prediction is made based on range of permittivity values. This project was totally new and it appeared most challenging to make data-set for predicting pure liquids range of capacitance apart from coming up with a device which can be easily used by all was itself a challenge. Got to learn how to think from low investment perspective to make efficient device easily affordable for common people.