May 2021 – Jul 2021 · 2 mos · Pune, Maharashtra, India

• Confidential Work

Apr 2020 – May 2021 · 1 yr 1 mo

• Dedicated 2000+ hours to a tenure comprising of all 4 tiers of hierarchy to India’s largest student-run NGO, Gopali Youth Welfare Society
• Leadership
• Increased the efficiency by 15% via collaboration tools & automation of tasks while leading a team of 50+ IITian’s
• Surveyed 1000+ villagers to assess their living standards in 27 villages with a team of 70+ members & 11 teachers
• Initiatives
• Coined JVMCare, a new fund-raising program to provide transportation to students and stability to teachers’ salary
• Transformed fundraising by introducing corporate matching which resulted in 1.2X higher per capita donation
• Co-initiated Hostel Construction for 150 students of classes 4-12 by successfully raising funds of INR 17L+ in 1 year
• Impact and Achievements
• Raised a total amount of INR 71L+ and saw YoY growth of ~17% in the financial year 20-21 amid global pandemic
• Provided free of cost education to 250 underprivileged students; generated permanent employment for 25 people
• Lauded by the Hon’ble HRM Dr. Ramesh Pokhriyal, & got featured in NDTV & The Telegraph(COVID relief initiative)

Apr 2020 – Apr 2021 · 1 yr

• Modelled a final algorithm for machine learning, software development, finance, and marketing to predict scores based on the resume.
• Preprocessed the project data using nltk, SpaCy, and Stanford CoreNLP Lemmatization and extracted the keywords from each project.
• Analysed the performance of a transformer-based model: Bert, XLNet (both pre-trained, fine-tune with relevant data) for generating word embedding and then multiplying with TF-IDF weights to generate sentence embeddings and Sentence BERT for direct generation.
• Analysed the performance of models like K-means, deep NN like Deep Embedded Clustering, Agglomerative, and Birch for clustering.
• Manoeuvred BS4, Selenium, scrapy, and a host of APIs like PhantomBuster for scraping 4L+ Google scholar profiles, 50k+ online courses, 3k+ LinkedIn profiles, 20k+ colleges from college Duniya and 30k+ companies, etc. for training and building the model.

Jul 2019 – Nov 2019 · 4 mos · Kharagpur Area, India

• Development Platforms Frontend: Python with PyQt4 for GUI,
• Backend: C++ with standard headers.
• Working with a team to develop software in which one can draw a graph with the user-friendly environment, move/delete edges and vertices, apply algorithms like Minimum Spanning Tree, blossom, BFS, DFS and other standard algorithms with graph data structures.
• UML Diagrams for a better understanding of the project and maintain the proper architecture.
• The final aim of the project is to create an application that is user-friendly and one can use all these algorithms without actually coding them.

May 2019 – Jul 2019 · 2 mos · Hyderabad Area, India

• Restricted Boltzmann Machine& ML classification algorithms to built a Deep Learning Architecture and predict Credit Card Defaulter.
• Achieved an F1 score of 34 percent and a sensitivity of 61 percent using (RBM + SVM).
• A probabilistic approach of Restricted Boltzmann Machine abstraction for feature selection and then processed the data with SVM, Naive Bayes, Logistic Regression, Decision Tree, MLP, etc. to predict the defaulter.
• Used the modified form of RBM such as Discriminative RBM and Deep RBM for improvements.