Developer Intern Data Analyst

Ayush Tripathi

LinkedIn: https://www.linkedin.com/in/ayushtripathi09/

Github: https://github.com/ayush0904

Email: ******@***.***

Phone: +1-213-***-****

EDUCATION University of Southern California, Viterbi School of Engineering, Los Angeles, CA Aug 2021–May 2023 Pursing Master of Science in Computer Science

Courses: Analysis Of Algorithms, Artificial Intelligence, Machine Learning, Web Technology, Database, Natural Language Processing. UP Technical University, Uttar Pradesh, India Jul 2013–May 2017 Bachelor of Technology in Computer Science and Engineering. Scored 81% with 1st division honors SKILLS Languages and Tools: Python, Django, Postgres, NodeJS, ReactJS, Angular, Flask, JavaScript, HTML, CSS, Bootstrap Libraries: AWS, Kubernetes, Kubeflow, Scikit-learn, Docker, GIT, JIRA, AWS, Tableau, MySQL, Cloudfront, MediaConvert Technical Skills: ML(Classification and Regression), Deep Learning (CNN, RNN), NLP(Bert, RoBERTA), TenserFlow, Pytorch PROFESSIONAL EXPERIENCE Software Developer Intern, Iteris Inc. May 2022- Present

• Designed and developed a Video Management Software using Node JS, and ReactJS. Formed pipeline to create video snippets and transcode videos with AWS ElementalMediaConvert, and deliver videos to client side with AWS CloudFront.

• Determined tech direction for ingesting videos. Created RTSP Stream server to receive 100+ video streams with FFMPEG, and store 300+ gigabytes of data daily in AWS S3 at a configurable interval.

• Developed restful service to upload files from RTSP stream server. Deployed the app using Docker Container and Nginx.

• Working on developing a 3 x 3 and 9 x 9 grid view video player program which support RTSP streams using Java and JFX. Graduate Research Assistant, University of Southern California Jan 2022- Present

• Created pipeline to execute Expected Maximization Algorithm to merge several datasets on basis of posterior probability. Extracted information from different Api’s. Successfully matched 80% of the regions(440/548) with FDR rate less than 5%. Gained experience on working with high performance computer.

• Applied Topic Modelling via LDA and Topic Bert (Model Coherence = 0.54) on corpus of 1.6 million tweets which is extracted from twitter using Tweepy Search Api. Experimenting various Machine Learning Classification Algorithms like XGBoost, RandomForest, and Bert to train on this corpus to identify general climate change misinformation. Data Analyst, DigiOne Technology Pvt Ltd Jan 2019–Jan 2020

• Developed assortment strategy for a retail store in India based on sales, exclusivity, and customer loyalty. Investigated clustering algorithms such as mini batch K means++ for defining assortments.

• Applied Market Basket Analysis to study association between products. Explored unsupervised learning algorithms such as APRIORI to study Association Rule Mining. Observed a gradual increase in sales up to 5%. Software Developer, Xceedance Technology Pvt Ltd July 2017-Dec 2018

• Designed one login customer portal with Apex(Java),Bootstrap, VisualForce and JavaScript for Insurance Major.

• Derived a solution to provide real-time integration between Salesforce and JDEdwards through a web application developed using RESTful APIs, Lightning Design System and Apex(Java) Language. PROJECTS Mini Go Game Feb 2022 – Mar 2022

• Designed and developed an AI agent to play a 5x5 Go game based on search game playing & reinforcement techniques. It was able to beat Random, Greedy Aggressive & Alpha-Beta player achieving an accuracy of 95%.GitHub Cipher Text Classification Jan 2022 – Feb 2022

• Implemented classification of ciphertext using various deep learning models like LinearSVC + TF-IDF character ngram(92.3%), fasttext + multinomial logistic regression(90.3%), Bag of n-grams adjusted using TF-IDF + ensemble of Naive Bayes and SVM(89.7%), and Bidirectional LSTM with FastText Embeddings(88.3%). GitHub Named Entity Recognition (NER) for Clinical Notes Aug 2021 - Oct 2021

• Identified and annotated specific clinical concepts in English patient notes using an ensemble model of Conditional Random Fields, LSTMs, and Transformer RoBERTa, achieving an accuracy of 88.2%.GitHub Stock Search (CSCI-571 Web Technologies) Jan 2022 – May 2022

• Delivered Website that allows users to search stocks using FinnhubAPI and display results on search page. Developed using NodeJS and Angular and deployed using Google App Engine. Created an android application for the same task. RESEARCH PAPERS AND CERTIFICATES

• Prediction of IPL matches using Machine Learning while tackling ambiguity in results (Publication) Oct 2020

• Air pollution in four Indian cities during the Covid-19 pandemic (Publication) Dec 2020

Contact this candidate