Ananya Rana

With a degree from one of the premier engineering institutions in India, IIT Kanpur, I am an ideal candidate for any Machine Learning and Data Science position. I have a solid technical background in various programming languages such as Python, C and Java. My expertise in Deep Learning, Computer Vision and Natural Language Processing makes me a valuable asset for any ML and DS related roles. I am passionate about using my knowledge and experience to identify and solve data-driven problems. I am confident that I can make a strong contribution to any organization.

• CS50 - Python, Harvard (edX)
• Machine Learning Specialization, Andrew Ng, deeplearning.ai
• Deep Learning Specialization, Andrew Ng, deeplearning.ai
• Intro to Machine Learning (STAT451), University of Wisconsin
• Intro to Deep Learning and Generative Models (STAT453) , University of Wisconsin
• Data Structures and Algorithms (Java), Imtiaz Ahmed, Udemy
• CS231n (2016) (Andrej Karpathy) - Computer Vision - Stanford University
• CS224n (2021) (Christopher Manning)–Natural Language Processing (ongoing)

Personality Type Prediction (NLP)

• Predicting the personality of a person from their last 50 social media posts.
• Used spacy library for text preprocessing, used various word embedding approaches such Tf-IDF, Bag of Words and n-grams. Realizing that the data was hugely imbalanced, I tried using SMOTE to address the issue as well. Then I used Word2Vec which generally works better with SMOTE, but in this case it didn’t. I have provided the possible reasons as to why it could have happened. Finally applied ensemble methods for prediction.
• Project repo link - https://github.com/ananya-rana/Personality_prediction.git

Neural Style Transfer – Computer Vision, Deep Learning

• This project uses a Convolutional neural network, a pretrained VGG-19, to capture the essence of both the content and style of two images and then combine them to create a new image.
• Because of low computing resources, I did not begin training from random noise vector, but instead began training by taking the generated image as a clone of the content image and made the generated image's parameters trainable so that it can learn the style of the style image.
• Project repo link - https://github.com/ananya-rana/Neural_Style_Transfer.git

Extractive Text Summarization (NLP)

• The algorithm measures the frequency of words in the document and decides the importance of the sentence considering the words in the sentence and their frequencies.
• Project repo link - https://github.com/ananya-rana/Extractive-Text-Summarization.git

Generative Adversarial Networks – Computer Vision, Deep Learning

• Using Deep Convolutional GANs to sample new Anime Faces . Since this problem has already been attempted countless times, I wanted to gain experience in working with complex neural network architectures such as GANs.
• Project repo link - https://github.com/ananya-rana/Generative-Adversarial-Networks.git

Car Insurance Claim Prediction – Machine Learning

• Performed EDA, found the dataset to be extremely imbalanced. Since such datasets may occur in the real world I wanted to work on such a problem.
• I applied Variance Inflation Factor and found the independent variables to be highly correlated, therefore I took steps to eliminate multicollinearity as well.
• Used imblearn library’s SMOTE technique to artificially create data to balance the dataset, and then applied ML algorithms. I also applied SMOTE’s relatives to the problem. However even after Hyperparameter tuning I did not achieve satisfactory results. I have provided the reasons why this could have happened.
• Project repo link - https://github.com/ananya-rana/Car-Insurance-Claim.git

House Price Prediction – Advanced Regression problem

• The dataset comprised of 80 columns, I performed EDA, applied end-to-end ML pipeline to predict the “SalePrice”. Achieved 86% r2score accuracy.
• Project repo link - https://github.com/ananya-rana/House-Saleprice-Prediction.git

Participated in ROBOCON 2015, All India Robotics competition

• Objective- to score points in a game of badminton which was supposed to be played by two robots of each team. The two robots had to serve or hit back the shuttle successfully to earn points.
• As a part of the mechanical sub-group, helped design and finally assemble the body, the working mechanism, as well as the badminton playing mechanism of the robot. The robot’s body was handcrafted through extensive metallurgical processes including work on the lathe machine, water-jet machine amongst other tools.
• Our team was placed 11th in the competition nationally.

Team Captain of The Mughals @Takneek, IIT Kanpur

• Led the team at the intra IIT Tech fest, comprised of miscellaneous competitions including in the field of Robotics, Civil Engineering (DESCON), Aeromodelling, Astronomy, Quizzes etc.

Secretary, Civil Engineering Society (The DESCON Club)

• Managed Club affairs, conducted Club activities in IIT.

Runners Up in Model United Nations @Antaragni Cultfest, IITKanpur

• As part of the Live crisis committee proposed solutions to the Rohingya crisis as it was being developed internationally. Donned the role of The Kingdom of Saudi Arabia and played mediator between the Islamic nations and the West. I was able to win the entire committee over by using diplomacy and got my way in almost all conflicts. Was given special appreciation by the chairperson for the role that I played in the committee.

Participated in Mr.Galaxy competition @Galaxy, IITK

• A competition held between individuals who are proficient in 6 fields – music, dance, drama, literature, arts and quizzing. Represented a 4th of IITK’s students as part of the Mughals team in the intra-IIT competition.

Won The Junior Saharanpur Cup in the District Level Cricket Tournament

Achieved 16th rank in Table Tennis UP State cluster competition

Runners up in District Classical Vocals competition

Graduated 4th Year(Senior Diploma) in Classical Vocals, 4th year(Senior Diploma) in Tabla, 3rd year in Violin and 1st Year in Bhaav Sangeet from Prayag Sangit Samiti, Allahabad.