Sonali Srijan

Email: adkc0e@r.postjobfree.com Phone: +1-765-***-****

AI enthusiast with experience in machine/deep learning. Seeking machine learning/software engineering internship positions for Summer/Fall ’21

EDUCATION Purdue University, West Lafayette, IN Expected graduation: 2022 M.S. in Computer Science & Statistics; GPA: 3.7

EXPERIENCE Ford Motor Company

Student AI Consultant, Natural Language Processing Aug-Jan 2021 As head of the Natural Language Interface to Databases (NLIDB) student team at Ford, my goal to develop and integrate a competitive deep learning based semantic parsing model into Ford’s pipeline for the conversion of user-generated Natural Language questions to SQL queries. A central obstacle to integration was the inability of most NL-to-SQL models to predict condition values in the generated SQL query. I successfully:

• Ideated an intuitive downstream Copy-Paste mechanism to predict condition values for RAT-SQL and other similar semantic parsing deep-learning models

• Trained a BERT-based deep learning model for condition value-prediction idea in NL2SQL models – Achieved an accuracy of 68% on RAT-SQL output and Spider dataset

• Integrated the trained BERT-based deep learning model downstream of RAT-SQL model for semantic parsing pipeline Purdue University, West Lafayette, IN

Graduate Research Assistantship, Bioinformatics Jun 2020-Ongoing PI: Dr. Kranthi Varala, Asst Professor

• Developed machine learning ensemble of SVM, XGB, KNN, Linear models to predict organ-type of plant using RNA- Seq data

• Built pipeline for quality control, preprocessing, downstream analyses of large-scale genomic data

• Used time-series based machine learning methods to infer Gene Regulatory Networks RELEVANT

PROJECTS

Meme Caption Generation using Deep Learning

Graduate Course Project, Natural Language Processing

• Developed an end-to-end encoder-decoder network (CNN + LSTM) for generating meme captions using Pytorch

• Developed a humor classifier neural network for memes Deep Learning based Defect Prediction Model for Source Code Graduate Course Project, Software Engineering

• Built a line-level defect prediction model for software source code from scratch

• Used BERT model to capture greater context of clean and buggy code. Distracted Driver Classification

Graduate Course Project, Data Mining

• Developed deep learning framework (CNN) for detecting distracted driver using Keras

• Leveraged transfer-learning methods (VGG-16, ResNet) for multi-class image classification Sentiment Analysis of European News Headlines

Graduate Course Project, Statistical Machine Learning

• Trained and tuned RNN, SVM and Random Forests for sentiment prediction of news headlines from Irish News Dataset on Kaggle

TECHNICAL

SKILLS

Python, Bash, SQL, R, Java, Git, Tableau, Web Scraping, NLP, Computer Vision Machine/Deep Learning: Pytorch, TensorFlow, Pandas, Scikit learn, HuggingFace, OpenCV, SQLite, Plotly GRADUATE

COURSES

Algorithms Analysis, Design and Implementation; Data Mining; Statistical Machine Learning; Natural Language Processing; Software Engineering; Networks and Communication AWARDS Honda Young Engineer and Scientist (Y-E-S) Award, India, 2017 Research Idea Presentation Award at Cognizance, IITR 2017 Youth Ambassador at JENESYS, Japan, 2016

Contact this candidate