Full-stack Architect and Developer in New York, NY, United States

The bot buys NFTs that it finds underpriced and resells them for a higher price. The project consisted of a few pieces.

Smart Contract
• A proxy contract has been created to allow for easy upgradeability.
• Role-based authentication is used so multiple accounts can interact with the contract
• Sends and receives ERC20 and ERC 721 tokens
• Interacts with contracts from NFT marketplaces to buy and sell NFS
• Deployed using Hardhat

Custom Harmony Node
• The bot searches for opportunities on the Harmony blockchain.
• The Harmony blockchain was forked, along with go-libp2p, which is used for networking. Heavy modifications were made so that the bot could listen to transactions in the mempool and submit the buy transactions in the same block.
• The custom node is deployed to multiple servers worldwide to reduce latency.

Searcher and Prediction
• The custom Harmony node uses precomputed price predictions based on different traits of NFTs.
• A Python application pulls data directly from the blockchain and the graph APIs, floor prices are computed, as well as a predictive model. These prices are then stored in Redis, which the custom node can access.

The platform helps data scientists quickly develop and deploy machine learning models for predicting user behavior. The system also allows data scientists to build dashboards quickly.

I developed the back end with Django and PostgreSQL, and the front end was built with Angular and TypeScript. Currently, it is running on Kubernetes on AWS.

The front end was built on Backbone.js, and the back end was built on Ruby on Rails, Clojure, Scala, Java, and PostgreSQL, along with featuring multiple services.

I was the first software developer at the company and later became the VP of engineering as the company grew to millions in revenue and over 100 employees. I architected the entire platform, developed it, managed the engineering team, and worked with Product to bring requested features to life.

The arbitrage bot found arbitrage opportunities across Curve, Balancer, and Uniswap and used Aave flash loans if possible.

The smart contract code to perform the arbitrage was written in assembly to reduce gas costs.

The bot was deployed on Ethereum, and a modified Geth node was used to allow the bot to listen for transactions in the mempool and submit arbitrage transactions in the same block.

The site allows the sale of locked tokens for the NFT game Defikingdoms.com. The front end is written in React, web3.js. A user can transfer his tokens to the escrow contract and set a price for which they can be bought. If someone buys the escrowed tokens, they are transferred to the buyer's wallet, and the purchase tokens are transferred to the seller's wallet.

Created the ERC20 Token with the following properties:
• It had a configurable percentage that was taken from all token transfers and rewarded to all token holders. This was implemented using reflection, as implementing this with a loop would be unfeasible because of gas constraints.
• It had a configurable percentage that was taken from all token transfers and deposited into a Uniswap type liquidity pool.
• It allowed for addresses to be blacklisted, ownership to be transferred, and tokens to be burned.
• The project was set up using Hardhat. Extensive tests were written and the Uniswap features were tested using a mainnet fork.
• The project was configured to be deployable to Polygon Testnet and Mainnet