• Led the development of pooled credit line and lender pool contracts, where lenders (either singly or in group) provide under-collateralized loans to a verified borrower.• Solicited interest from 5 lenders (20% of total) through Discord events, rallying 50+ stakeholders to join.• Achieved 100% security compliance by undergoing a Web3 audit amid rising security attacks.

• In an attempt to start up, tried several ideas across different domains (Details: https://learned-ping-1cd.notion.site/Starting-Up-8fee0b3124c54daea7311707a6278cf8).• BNPL payments at offline stores via QR: Did market research, built pitch deck, demo video & applied to YC, left it due to concerns regarding revenue and regulations.• Health social commerce platform: Talked to ∼50 dermatologists and nutritionists for partnerships, built an initial MVP mobile application, left it due to lack of FMF.• Marketing Automation for DTC brands: Talked to ∼100 PMs to understand domain, decided to focus on D2C brands as GTM strategy, reached out to ∼40 DTC founders.

• Worked as a risk modeler for Interest Rate asset class in the Market Risk division.• Managed ownership of several regulatory models, which were used for internal purposes and for fulfilling the requirements of Fed, Basel, European Union.• Developed new IR stress test market scenarios to examine vulnerabilities of berms, exotic vol products to rates and volatility movements.

• Shifted Data Retriever's JSON script metadata to a separate project location to help with maintenance, and testing.• Worked on overall improvements in the core application, and to enable using the latest version of JSON scripts metadata.• Worked on the wrapper repositories based on R, and Julia to ensure compatibility with the new design changes.

• Interned as a Software Developer under the Amazon Fresh - Category Management group.• Worked on the task of automatic synthesis of a large number of pipelines using Live Pipeline Templates (LPT) scripting.• Made necessary design changes in the React Native and Spring MVC based frontend and backend applications.

• Worked towards improving the accuracy of 3D human pose estimation for images of squats and conditioning exercises.• Read some research papers and performed re-training of a couple of them on images of strengthening exercises.• Used PyTorch and TensorFlow as the software packages for all the experiments.

• Worked on the automatic colorization of low-resolution greyscale images using a Conditional DC-GAN.• Used L2-Regularization along with Adversarial Loss to train the model which removed brown-ish hue and green patches.• Utilized L*a*b* color space for minimization of color and brightness fluctuations in the resulting image.