Resume PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

Paritosh Pankaj

Third Year Undergraduate, Statistics and Data Science # [email protected] ‡ +91-9153226981


Indian Institute of Technology Kanpur § ParitoshP702 ï Paritosh Pankaj

Academic Qualifications
Year Degree/Certificate Institute CPI/%
2021 - Present BS(SDS) Indian Institute of Technology Kanpur 8.8/10
2021 CBSE (XII) New Horizon School, Bhagalpur 96.4 %
2019 ICSE (X) Mount Assisi School, Bhagalpur 96.6 %

Publications
• ‘A Linear Programming enhanced Genetic Algorithm for hyperparameter tuning in machine learning’, Ankur
Sinha,Paritosh Pankaj, accepted to be published in IEEE CEC 2023.
Internships
Reinforcement Learning for Dynamic Pricing| Reasearch Intern| IIIT Hyderabad| Prof. Tejas Bodas(May’23 - July’23)
• Explored different RL algorithms for dynamic pricing of products for both deterministic and stochastic nature of demand.
• Implemented GILS,CILS, Parameter Space Exploration and Thompson Sampling for pricing in infinite inventory.
• Applied REINFORCE, Natural Policy Gradient,A2C,TRPO,PPO and DDPG for pricing in finite inventory settings.
• Forged a fast converging data efficient low budget Bayesian optimization algorithm for searching directly over policy
parameters which beats the control based state-of-the-art Reinforcement learning algorithms used for dynamic pricing.
Data driven Market Segmentation| Machine Learning Intern| FeyNN Labs (Apr’23 - Jun’23)
• Conducted a study on different layers and approaches of multidimensional Market Segmentation analysis.
• Examined and rendered different algorithms like K means Clustering, GMMs and Agglomerative Clustering.
• Deployed these clustering algorithms to find the states in India suitable for establishing an Electronic Vehicle startup.
• Used the same approach to build a small scale model for predicting which kind of people are more prone to heartstroke.
Machine Learning using Bilevel Optimization| Reasearch Intern| IIM Ahmedabad| Prof. Ankur Sinha(Dec’22 - Jan’23)
• Formulated the hyperparameter tuning problem in Machine Learning as a relaxed bilevel program.
• Tuned the model hyperparameters using a micro Genetic Algorithm and solved a linear program of 14000 variables.
• Improved upon the regularization penalties of the multi-layer perceptron model to achieve an accuracy of over 82%.
• Developed the algorithm which beats the other baseline methods of hyperparameter tuning by 10 folds.
Key Projects
Big Data Visual Analytics| Supervisor: Prof Soumya Dutta (Oct’22 - May’23 )
• Reviewed advanced, probabilistic data-driven methods of sampling 3 dimensional, time-varying mulitvariate datasets.
• Executed Gradient based sampling, value based importance sampling and pointwise mutual information sampling.
• Cited different measures of Specific Information like surprise(I1), predictability(I2) and entanglement(I3).
• Implemented and analyzed the use of these measures for transfer function design to enable better data visualization.
Skim Lit Pub Med| Self Project (Oct 2022 )
• Implemented the research paper which talks about the sequential sentence classification of medical abstracts.
• Created a baseline model by pipelining the tfidfVectorizer and Multinomial Naive Bayes Layer.
• Constructed a Conv1D model using token embeddings and enhanced the performance using pre-trained embeddings of Uni-
versal Sentence Encoder to achieve an accuracy of 83% along with a f1 score of 0.8317.
Web scrapping and Data Analysis in R| MTH208 course project | Prof. Dootika Vats (Sep’22 - Nov’22 )
• Scraped the data of top 100 influencer channels on youtube from the official site of Noxinfluencer using tidyverse library.
• Extracted information like noxscore, total video time, total subscribers on youtube, total number of views, language used,
area of operation. Analyzed the data graphically using ggplot and by creating an interactive Shiny app using RShiny library.
Technical Skills
• Programming Languages & Utilities: C, C++, Python, R, SQL, Git, LATEX
• ML resources & Libraries: NumPy, Pandas, Scikit-learn, TensorFlow, PyTorch, Imager, Rvest, RShiny, Tidyverse
Achievements
• Received the Academic Excellence Award for exceptional academic performance for the academic year 2021-22.
• Secured All India Rank 4671 in JEE Advanced 2021 among 150,000+ shortlisted candidates.
• Among 458 students to qualify for INPhO 2019 conducted by HBCSE out of about 54000 appearing candidates.
• Awarded A grade for outstanding performance in Theory of Probability and 9 other courses at IIT Kanpur.
• Participated in National level of 25th National Children’s Science Congress 2017 organised by NCSTC.
Relevant Courses
A: Outstanding Performance
ESC101: Fundamentals of Computing (A) ESO207: Data Structure and Algorithms MSO205: Theory of Probability (A)
MTH210: Statistical Computing (A) MTH211: Theory of Statistics MTH208: Data Science Lab
MTH102: Linear Algebra and ODE (A) MTH212: Elementary Stochastic Process MTH207:Matrix Algebra Linear Estimation
MTH209: Data Science Lab 2 MSO202: Complex Analysis (A) MTH301: Advanced Real Analysis

You might also like