EQUI SWAP
Aggregator/Swaping tool/Wallet manager
Pitch: Equi Swap - Pitch 2025
GitHub repo: EQUI SWAP
We use our own indexer with microservice architecture. Each user is handled in a new thread with indexer as a separate service
Tech diagram
Recent updates:
AI assistant ESP32
Small (6x3cm) portable device which you can talk to while holding the pressure sensor and get a response on display
Device connects to a phone with BLE and requires a simple android app which works as a communicator between device and server, so you dont need to connect this device to a wifi or buy separate sim card
Solana Arb
GitHub repo: PRIVATE FOR A REASON
Software works by having indexer which monitors blockchain calculates prices every time it receives an update from gRPC.
Software has its own simulation service so there is no need to make any API calls which minimizes the delay.
Deployed program recalculates the given transaction to check if any changes happen during transaction landing.
Made in rust to maximize performance.
Show More (old updates)
(Q2 2024) At the moment works by retrieving data from self deployed Jupiter API (we as a team want to move away from this concept and get token prices on our own). Our bot is better than most known shared bots, by performance and bringing new features.
Why is our bot better?
threads threads threads ram ram ram...
Worker config:
Lets user define different wallets to be used by different workers.
Lets user define max transaction age (by default if solana transaction is not landed in 150 slots its being failed. Solana doesnt provide overriding MaxTxAge variable, but we still did it.).
fetchInstructionTimeout allows your bot to stop executing a potentially profitable transaction if it took too long to fetch instructions...
No need to look for tokens and update includeMints.json as bot does it for you.
A lot of performance enhances, reducing time before transaction execution and even time to land transaction. (No details for a reason)
SEEFOOD
A mobile app that uses AI to detect whether a given food item is a hotdog or not a hotdog. Originally a fun parody project, it demonstrates skills in machine learning, image classification, and mobile development.
UI mimics the original app from the HBO series Silicon Valley
Library ERP
The main idea of this project is to make a GUI and web site application which will be able to work with SQL databases, will be fully customizeble to use it for any needs.
GitHub repo: LibraryERP
I tried to make the whole application with mvc design pattern so its easy to make changes and make updates to it.
The whole functionality available at the moment is represented on the diagram bellow.
Recent updates: