Sandeep Yadav
+91-735******* linkedin.com/in/sandeep
admr9j@r.postjobfree.com github.com/sandeep021
EDUCATION
Undergraduate Institute of Engineering and Technology, Lucknow Aug 2017 – May 2021 B.E- Computer Science and Engineering CGPA: 8.14
Experience
Intern CETPA INFOTECH Jun 2019 – July 2019
• On breast cancer data (size: 369 instances) applied instance-based learning algorithm (KNN). Collected classification results averaged over 10 trials best accuracy results on 1-nearest neighbour (>92.7%) trained on 200 instances, tested on the other 169.
• Made a UI that can detect a phishing website and will warn you, used logistic regression for classification (accuracy> 83%).
• Build a python UI that predicts next 20 days live stock prices of organisations using linear regression (accuracy>99%).
• Tech: Python Scikit-learn pandas NumPy tkinter. Technical Projects
ISRO Bhuvan Portal’s Products Review Analysis.
• Sentiment Analyser with Queue-Work Architecture implemented using Celery for an asynchronous task queue based on distributed message passing and Redis as the message broker.
• Achieved 93% accuracy using CNN architecture on word vocabulary matrix of size 100k, Used Spacy for data pre-processing.
• Used Glove for word embedding (converted words into Tensors of size 300). Tested it for 77k reviews under 300sec of CPU.
• Tech: Django Celery Redis PostgreSQL Pytorch Glove Spacy Pandas JavaScript.
Twitter Tweets Sentiment Analysis in Real Time.
• A python application which analyse & predicts the sentiment of tweets of entered twitter handle in numerical, graphical forms and also in emoji expressions tweets data is fetched using tweepy API.
• The unwanted characters which don’t play role in deciding sentiments are removed initially to reduce computation time and prediction using re python module.
• Improved accuracy by word tokenization, lemmatization and agumentation.
• Tech: Python tweepy tkinter TextBlob re (regular expression).
Web Portal Where You Can Put your Secrets.
• A secured web application that allows you to login and then post you secrets, you can also view other secrets anonymously.
• Implemented encryption, Hash function encryption & salting + hashing to secure the database hosted on AWS cloud.
• Cookies are added using passport to save the recent sessions in browser for quick authentication and other browsing details.
• Introduced o-auth for getting third party (google, Facebook) login & signup ability to achieve their security features.
• Tech: NodeJS Express passport EJS HTML CSS Bootstrap JavaScript MongoDB. ACHIEVEMENTS
• Solved about 800+ coding question on various online coding platforms (leetcode codeforces codechef spoj).
• Ranked 761 out of 9165 participants in Google Kickstart Round H 2020.
• Kaggle contest rank under 80 out of 4k participants for my CNN architecture model.
• Smart India Hackathon 2020 Finalist for the ISRO’s Problem Statement, our team is in Top 3(out of 279) teams at national level.
• Mentored students of different colleges of India at IIT Kharagpur open source competition.
• Completed NPTEL course on Machine learning for engineering and science application (scored 67% marks). (certificate).
• 65+ GitHub contributions, holds badge of Arctic Code Vault Contributor. SKILLS
Strengths: Data Structure Algorithms Machine Learning Backend Development.
Languages: C++ Python.
Database: MongoDB, MySQL
Technologies: NodeJS ReactJS Express sckit-learn NumPy.
Environment and Tools: Linux Windows Git
Other & Hobbies: Leader Teamwork Cricket Gym.