Rishi Suvvada

[email protected] | linkedin.com/in/rishi-suvvada/ | github.com/Rishi9247

Education
Indian Institute Of Technology Kharagpur,Kharagpur CGPA-9.19/10
Interdisciplinary Dual Degree- B.Tech in Chemical Engineering, M.Tech in Artificial Intelligence July 2019 – Present
Sri Chaitanya College,Visakhapatnam CGPA 10/10
Andhra Pradesh Board of Intermediate Education(XII) June 2017 – May 2019
Sri Chaitanya School,Visakhapatnam CGPA 10/10
Central Board of Secondary Education, CBSE(X) April 2007 – March 2017

Internships and Projects

Meta-Adapting Hilbert Filter EEG classification | Under Prof. Debasis Samanta Aug’22 – present
• Implemented Transfer learning for lower data Interpretability of EEG signals using the Meta-learning approach for

extracting subject-invariant representations of data.

• Performed data-driven frequency filter transformation (Hilbert Transform inspired from HTnet) for initial data to

boost generalization of cross-modality learning of the model.

• Achieved accuracy of 77% while testing on BCI-IV 2a EEG dataset of 4 class motor imagery dataset. Hilbert

transform layer improved the accuracy of prediction by 2%

• Testing out our model on other meta learning public datasets like DEAP dataset for analyzing the impact of

Hilbert transformation layer on the meta-transfer learning technique,Intending a paper publication from analyzing
the results from both the datasets
EarlySalary - Instant Cash Loan, Pune | Data Scientist Intern May’22 - July’22
• Developed a Score-card for new-to-credit(NTC) customers using look-alike thick-file population (existing credit

borrowers) for better data and risk analysis

• Performed K-Prototypes Clustering to find look-alike of NTC in thick-file, leveraging the population of thick to

train our model, Implemented Weight of Evidence (WOE) binning with Information value (IV) and Cramer’s
threshold as a metric to filter good features
• Analyzed default rate curves w.r.t WOE-binning for each variable to make business insightful variable tweaking and

selection and Performed Logistic regression on the following features to achieve a AUC of 0.72
• Achieved a swap-in rate of 24 percent on out-of-time testing with uniform population distribution and steep Risk

rankings
C-Language Parser | IIT Kharagpur Sept’21 – Nov’21
• Building custom made, optimized BNF format grammar rules for parsing C language program

• Performed lexical analysis using the lex module from the open source ply library and tokenized the C-code into

terminal tokens using regular expressions from regex module

• Developing a parser to perform snyntatic analysis using the BNF grammar build earlier with the help of yacc

module from the ply library

• Constructed the abstract syntax tree (AST) structures for the parsed code for better visualization.

Computer Vision based COVID safety surveillance | Under Prof. Adway Mitra Nov’20 – Feb’21
• Developed a model to predict the spread of COVID-19 in an area.The model takes information such as the

population of the area, the number of people for which data was taken,number of people who were wearing masks
versus the number of people who didn’t wear masks.
• Built a Convolutional Neural Network model to predict whether a person in the photograph is wearing a mask or

not.The data was labelled manually using LABELIMG opensource software and data was collected from images
captured in the tech market and other areas from IIT Kharagpur.
• Implemented YOLOv3 for Face and Mask Detection and achieved an accuracy of about 85 percentage.The accuracy

and speed of detection was further improved by using YOLOv5 which gave about 95 percentage accuracy.
• Attempting to deploy the overall system in public places like Administrative office and Tech Market insides the

campus.
National Artificial Intelligence Research Portal | Under Prof. Sudeshna Sarkar Nov’20 – Feb’21
• Developing a national portal for learning different algorithms involved in Machine Learning
 • Developed various ML assignments on Decision Trees,Ridge and Lasso Regression,Logistic Regression,CNN
• Evaluated the assignments automatically using an evaluator code in python with json being the intermediate.
Implemented the models from both scratch and libraries like ScikitLearn, Pandas, NumPy, Seaborn, etc
• Deployed the assignments on the NAIRP official website where students can code directly on the website with the
help of cloud-based GPU sponsored by Amazon(AWS) sage maker

Competitions
Hult Prize On-Campus 2022 | 1st runners-up March’22
• Identified a real life need of catering to the needs of working professionals to serve them with home style hygienic

meals
• Developed a scalable business model to monetize home cooked hygienic food with meaningful employment to

homemakers
Open IIT Data Analytics | Winners March’21 – April’21
• Supervised, Imbalanced classification of songs according to popularity for biding and earning maximum revenue

• Analyzed the Correlation matrix and Scattered plots trends of the acoustic features to draw major insights into the

dataset
• Compared accuracies of 5 different ML models like Random forest, SVM, Decision Trees,etc. with hyper parameter

tuning
• Integrated the above models into an Ensemble weighted voting classifier for maximizing revenue generation and

accuracy
• Achieved best results on Ensemble model with maximum revenue of $9136 and accuracy being the trade-off settling

at 73%
Kharagpur Data Science Hackathon 2021 | 3rd Runners Up Feb’22 – Mar’22
• Predicting the price leaders and followers amongst several bus services using time series data
• Performed EDA and feature engineering to draw major insights into the data set and created 5 new features for
better model performance
• Generalized the non-uniform time series data using SVR ( RBF kernel) and estimated price at uniform regular
intervals
• Implemented Agglomerative Clustering and Dynamic Time warping for measuring similarity between two temporal
sequences of bus pricings
• Integrated both into an Ensemble model and performed Granger Causality Tests to final predictions for finding
market leaders and followers accordingly

Coursework
Programming:Algorithms-1(CS21003) — Algorithms Lab (CS29003) — Deep Learning Foundation and
Applications(AI61002) —Natural Language Processing(CS60075)— Image Processing(CS40019)— Machine
learning Foundations and Applications(A142001) — Data Structures and Algorithms(Algozenith) — Data
Analytics(CS61061) — Artificial Intelligence Foundations and applications(AI61005)
Mathematics: Probability and Statistics(MA20104) — Linear Algebra for AI and ML (AI61003) —
Numerical Solutions of PDE (MA20102) — Transform Calculus(MA20101) — Maths-I(MA10001) —
Maths-II(MA10002)
Achievements
• Awarded AI-ML Interdisciplinary Dual Degree specialization for being in top 31 students in whole 3rd year
applicants- 2022
• Secured a departmental rank of 3 among the 70 Dual Degree students of the department of Chemical Engineering at
the end of 6th semester- 2022
• Stood in the top 2% of total 2 million applicants in JEE-Mains and top 1% in JEE-advanced- 2019

Skills and Expertise

• Programming languages : C, C++, Python, HTML/CSS, Prolog, SQL, MATLAB
• Libraries : STL, Pandas, NumPy, ScikitLearn, Matplotlib, Tensorflow, PyTorch, YOLO, Seaborn
• Software : MATLAB, GoogleColab, Jupyter notebook, LTspice, SolidWorks, MS Office, Adobe Photoshop, Figma