Cofounder & CEO of oneinfer.ai, Building openbandwidth the next frontier for accessing AI

youtube

devto

medium

productHunt

github

Achuth Kandikunnta  • Peerlist

Built an Arbitrage bot to track 5 decentralized exchanges present on the Ethereum blockchain (Uniswap,Sushiswap,Bancor,Kyberswap,Curve.fi). Bot listens to the swap/sync events emitted on the blockchain for every 14 sec(avg.) and calculates potential arbitrage opportunities after deducting txn. chrgs.. The Algorithm used to calculate the opportunities is based on graph theory. Then it creates the txns(capable of taking flash loans) and includes them in txn. bundle(using flashbots libs) to include in the next block. Tech Stack used: Python, nodejs, Redis, solidity.

DEX Arbitrage bot

In this project,News Event Exploration system has been designed for exploration and enrichment of news events and concepts. Although, there are few systems which provide up-to-date related news information in real time, but they overload the user with large amounts of results. Clearly, there is a need of a system, enabling readers to get a broad overview of news data generated in response of user query. In order to learn a global description of news while maintaining a fast and accurate search, we propose a keyphrase based news exploration system, Our system works in two  phases: keyphrase extraction followed by keyphrase enrichment. The keyphrase enrichment phase finds important and interesting data related to the keyphrase, such as connected entities; emerging, active news concepts; and roles of associated keyphrase. Also, Data from Dbpedia data sets is also used to generate related information like images, short abstract, long abstract etc.. about the the event. Also Images related to the event are scrapped from the news websites.

Event Exploration System

News Event Exploration system has been designed for exploration and enrichment of news events and concepts.Although, there are few systems which provide up-to-date related news information in real time, but they overload the user with large amounts of results. Clearly, there is a need of a system, enabling readers to get a broad overview of news data generated in response of user query. In order to learn a global description of news while maintaining a fast and accurate search, we propose a keyphrase based news exploration system, Our system works in two phases: keyphrase extraction followed by keyphrase enrichment. The keyphrase enrichment phase finds important and interesting data related to the keyphrase, such as connected entities; emerging, active news concepts; and roles of associated keyphrase.

Keyphrase based News Event Exploration Engine

Text line segmentation is an important step for OCR. If a document contains multi-oriented text lines,	proper line segmentation is	to be done. There are many methods for segmenting lines in normal text	documents. But if the document contains handwritten text written in different directions normal techniques can't be applied. So, we separated the the text in specific orientation and do line	segmentation for that separated text.