A web3 platform that enables creators to launch and grow their business brand with ease. Spur takes away cumbersome tasks like global payouts, invigorating loyal patronage through virtual/physical events and novel incentive programs. Spur engineered an NFT minting/gifting app that further saw the growth of its community.

Developed an Optical Character Recognition software for a non-profit to recognize regional scripts (Malayalam & Syriac) from a picture (.png) and generate its editable text equivalent for digital text publishing. This project facilitated an understanding of heuristics in decision making of an AI system. Feature extraction is done using a 7-layered convolutional network (CNN), which is then fed in to Bidirectional LSTM.

Full Stack Developer . Won 'Innovation Champion' award for engineering a payment reversal system for corrupted commission payments. Played a significant role in transforming a highly error prone Tax Reporting system into a Zero-Defect system within a year. This helped the company shelve an elaborate redesign effort. Conceived, designed & developed a Code traverser tool – IntelliPreter, capable of deepdiving several layers of programs to find potential infinite loops and initialization errors that can happen during an actual execution , due to bad coding standards.

Conceived, designed & developed a Code traverser tool – IntelliPreter, that could deep dive to any levels to bring out inconspicous but fatal infinite loops between a group of programs in coupling. Since it had the ability to alert about variables being used without resets [potential error causers] and the ability to optimize traversal , it was named IntelliPreter - an intelligent Interpreter. This was useful in tackling bad coding practices in vintage legacy codes. Was awarded the 'Innovation Champion' award.

Developed a Loyalty Card Wallet App. The Android app employs Computer Vision and Image Processing methods by making use of inbuilt camera to scan a Card with Barcode / QR code to identify the Format and Unique number [key]. The application captures the barcode as a bitmap in grayscale and then decodes the format and other meta data from its bitmatrix. The two attributes are then stored away in memory. When the user wants to retrieve the card details at a counter [e.g.: retail], the app would regenerate the Barcode (i.e. redraw the bitmap) from these two attributes in memory at runtime.