Sahil Sunda

Technical Intern

• Built the pipeline of PM pillar dashboard for real-time data insights of equipments.
• Configured the interface of Sandbox server using ABAP in SAP S/4 HANA for Costs Stratification, defined cost elements and their GL accounts, settlements, etc.
• Stratified the data-endpoints of RCA & Failure data to data lake, which later can be directly connected to Tableau Server for showing the trends.
• Executed all transactions of quality server of SAP, proposed the wireframe for Costs Stratification, RCA & Failure data for component-wise analysis to Accenture team.

Software Developer

• Building dynamic E-Commerce web-application from scratch with different open source integration’s to provide a platform to the community for processing, visualizing & exploring Geo-Data sets of Drone images, Satellite Images which they can sell.
• Using ExpressJS, Docker, MongoDB (for storing login details), PostgresSQL (for managing user data) at the backend, and AWS for the storage, and for frontend part ReactJs & for authentication, used JWT which would be changed before deploying.
• Part of the technical team of the firm, working on both service-side and customer-side to provide a platform for the processing & sharing of the data.

Software Developer

• Built an authorization-master API from scratch which includes signin, signup, authentication using JWT, Expressjs, Typescript, Mongoose, Docker for automation.
• Restructured code and designed the user-end of an online learning platform
• Worked to build an API like Commento using NodeJS, ReactJS to make it like a script which can be added to website and can be used for discussion.

Web Developer

• Structured & built new functionalities for the Website of the NGO to make it look more dynamic I had used Javascript, HTML, Bootstrap.
• Debug the code using Chrome Dev Tools and made the website more user friendly.
• Configured various other websites to make the relatable website for the NGO.

IIT Kanpur

Rajat Vidyapeeth School

Dundlod Public School

MICROSOFT CYBERSECURITY ENGAGE

• Learnt industry-relevant Cybersecurity skills from professionals of Microsoft.
• Practiced computer forensics, web-app penetration testing, ML for security, firewalls, SIEM, malware analysis, SQL injection, spoofing and cyber defence.
• Submitted a research paper on prevention of ”Attack on Electronics Systems of Nuclear Power Plant”, planned its methodology, solution architecture & benchmark for success.

STOCK PRICE PREDICTION

• Aim of the project was to use Machine Learning models to predict a company’s stock price.
• Employed Data visualisation to compare the Return vs Risk of investment, Daily Percentage Return and Moving Average trend over time for four companies from different sectors including IT, Banking and Petrochemical
• Used Augmented Dickey–Fuller Test to test the presence of stationarity and checked for seasonality
• Used SARIMAX, Linear Regression and LSTM and achieved lowest RMSE of 35.3 and highest r2 score of 68%

DATA ANALYSIS IN CRIMINAL JUSTICE

• Analysed FIR Data using linear, logistic regression models, pandas, matplotlib.
• Practiced Agent Based Modelling (ABM), Network Analysis for Law Enforcement (using Gephi), Visualisation of SATP (south asia terrorism portal).
• Worked on the final Research Project, planned its research methodology, literature review, Data Source & Data Analysis. Workflow includes SARS model.

IMAGE PROCESSING USING CROP IMAGES

• Squared the project with the collaboration of IITKGP Prof Dillip Swain and an agro-based Tech company KhetiPoint for the Crop images Data-sets.
• Extracted images from the raw Database(firebase) of the firm, then filtered and split the data randomly into test & train using python libraries and packages.
• Worked on Data Mining (using pandas, urllib), Data Splitting (using random, shutil), Data Augmentation (scipy, numpy, skimage), Data Visualisation (matplotlib).
• Finally implemented ML Model using Tensorflow, Keras, VGG-16 architecture for the image classification & detection.

COVID-19 TWEETS SENTIMENT ANALYSIS

• Aim of the project was to classify a dataset containing approximately 179,100 tweets with 13 different features as positive, negative or neutral.
• Data pre-processing included removing punctuations, emojis, stopwords, lemmatization, & was done using NLTK library
• Feature extraction was done using Bag-of-Words and Tfidf Vectorizer classified as positive, negative or neutral. Performed interactive data visualisation using Plotly and WordCloud.
• Gained an accuracy of 91% using Embedding layers and LSTM cells and BatchNormalization layer to avoid overfitting