Python Data

T e j a s w i n i K a n c h a r l a

Email : ******************@*********.*** Phone : +1-970-***-**** Github : tejaswini4 Linkedin : Tejaswini kancharla Program Institute Year Aggregate Score

Masters in Computer Science Colorado State University 2018 - 2020 3.76 B.Tech in Computer Science Sri Venkateswara University College Of Engineering 2014 - 2018 CGPA-8.0

• Programming Languages: Python, Java, GoLang, SQL, UNIX shell scripting, JavaScript, C/C++,

• Coursework: Machine Learning, Artificial Intelligence, Image Computation, Big Data, Object-oriented Design, Database Systems

• Frameworks: NLTK, Gensim, Spacy, Scikit, Keras, Tensorflow, Apache Spark, Hadoop, Apache Storm, NumPy, Pandas WORK EXPERIENCE

• Data Science Intern (Prediction Health, May’18 – Present, Manager: Pedro Teixeira, M.D., Ph.D.)

Topic Modelling: Built a model that would classify the content of the notes into different sections. To achieve this, it was required to analyze the MIMIC dataset using LDA and t-SNE.

Technologies Used: Gensim, Textmining, Scikit, TensorFlow, Python, TPU, Google Cloud Platform (AI Platform)

Speaker Prediction: Extended the use case of BERT by fine -tuning the BERT Large to predict the role of the speaker given the text. Eg:(doctor, patient). Deployed the model on the Google Cloud Platform using custom prediction. Technologies Used: TensorFlow, Python, TPU, Google Cloud Platform (AI Platform)

BigQuery API: Developed an API so that one can access the data stored in Big Query tables of Google cloud platform without having to worry about the schema of the table or SQL call.

Technologies Used: Python, Google Cloud Platform, Big Query

Data Visualization of Large Data: Visualized huge medical data with pie charts, bar graphs, histograms that gives a good idea about the most common symptoms, diseases, and medications etc., which help our customers to improve their business. Technologies Used: Big Query, Matplotlib, pyplot, Jupyter lab PROJECTS

• User Based RNN for Time Aware Movie Recommendation System Technologies Used: SparkFlow, HDFS, Java Implemented Time Aware Movie Recommendation System that extracts the patterns in the watch history and reviews of the people from big data and recommends new movie. We tried different architectures like Standard GRU, Standard LSTM, Linear User-based GRU and Regulated User-based GRU and achieved higher MRR@20(0.05) and Recall@20(0.19) with Regulated user-based GRU.

• Real-Time Streaming Data Analytics with Twitter Technologies Used: Apache Storm, Twitter API, Scala Implemented loss counting algorithm using Apache Storm that would execute in parallel to extract the top 100 popular topics from twitter live message streaming for every 10 seconds. Our implementation impressed the instructor as our accuracy was better than all the other teams she has seen from past few years.

• Animal Recognition Technologies Used: OpenCV, TensorFlow, Keras, Python Collected and preprocessed the dataset of 8 different animals during day and night. This dataset was then trained on by a CNN and used to pedict the animals. The accuracy was 99% as long as you use one of the 8 animals’ images and try to predict. If we give an animal which is not in the dataset then it would try to come up with the animal name that closely matches.

• Tracking Moving Objects Technologies Used: OpenCV, python Developed a program that would track the objects of interest. You can select an object in the screen by a drawing a bounding box around it using the mouse and then the program starts tracking it. I have used Mosse’s tracking algorithm to achieve this

• Speech Recognition Technologies Used: TensorFlow, Keras, SciPy, Google Colab Developed an LSTM Recurrent Neural Network on the Google dataset obtained from Kaggle Competition for speech recognition. The objective was to recognize 10 different words in the dataset. Our model has achieved an accuracy of 99%.

• One Shot Teaching Technologies Used: NLTK, Python, Files Developed an application that would help to customize the commands not through settings by actually using it. This application learns from user feedback. I just made a simple application that would perform addition and subtract ion and you can change your commands for doing these operations by giving the command to the application and saying whether it did a right or wrong thing.

• Chatbot WebApp Technologies Used: Django, HTML, CSS, Java Script, Speech to Text API and Grammarly API Developed a webapp that hosts chatbot which helps people to learn English. I have integrated Grammarly API into the chatbot so that it would take care of the corrections. The user can speak to the bot which will extract text from speech and show the corrections while asking questions to make the user speak more.

University Innovation Fellow of Stanford University and facilitated a number of design thinking workships

First Chief Operating Woman Officer of Ignite, an entrepreneurship club at Sri Venkateswara University College of Engineering. TECHNICAL SKILLS

EDUCATION

EXTRA-CURRICULAR ACTIVITIES

Contact this candidate