Explore my projects

TikitWorld

Tikitworld

Some of tikits features

TikitWorld

TikitWorld is a decentralized, tamper-proof ticketing platform that revolutionizes event management for owners, customers, and brands. It ensures fairness and security in the secondary market.

Skill Set & Technologies

FullStack
Blockchain
DevOps
RabbitMQ
Docker
REST api
Python/Django
Html/React
Postgresql
Solidity/Rust/Cairo/Motoko
2021 - 2024
CTO/Lead developer

Technical description

TikitWorld exemplifies the power of Blockchain, integrating multiple systems to create a robust platform. It operates on an on-chain SmartContract deployed across several chains, working with the backend to ensure decentralization and transparency. This setup mitigates adoption challenges by managing user wallets on the backend.

The platform uses a PostgreSQL database synchronized with the blockchain to maintain object states. An asynchronous service batches and checks transaction states, while full system containerization ensures seamless scalability.

Technical description

TikitWorld exemplifies the power of Blockchain, integrating multiple systems to create a robust platform. It operates on an on-chain SmartContract deployed across several chains, working with the backend to ensure decentralization and transparency. This setup mitigates adoption challenges by managing user wallets on the backend.

The platform uses a PostgreSQL database synchronized with the blockchain to maintain object states. An asynchronous service batches and checks transaction states, while full system containerization ensures seamless scalability.

Docos

DocOS is an AI-powered decision-making solution that transforms organizational documentation processes. It streamlines document digitization, analysis, collaboration, and decision-making.

Skill Set & Technologies

FullStack
AI & Machine Learning
DevOps
RabbitMQ
Docker
Python/Django
Html/React
Postgresql
Vector Database
Open/Elastic search
Redis
Postgresql
Websockets
Rest API
2023
CTO/Lead developer

Technical description

DocOS is one of the most complex projects I've developed, integrating multiple systems (pipelines):

Document Processing Pipeline

This pipeline analyzes uploaded documents, splitting them into semantically relevant chunks, and qualifying, summarizing, and extracting data. The documents are stored in three databases: PostgreSQL, Vectorized database (embeddings), and OpenSearch (or ElasticSearch). The pipeline also reviews existing documents for redundancy and conflicts.

Database Search Pipeline

This pipeline searches multiple databases using keywords and phrases, scoring each result. It employs "prompt engineering" to refine search terms for optimal database queries.

Multi-LLM Pipeline

The platform integrates multiple LLMs for various tasks. It supports local LLM deployments for privacy, with a default small local LLM for prompt engineering and large LLMs for chat-like experiences (e.g., GPT-4, Claude).

Prompt Engineering and WebSocket Pipeline

This layer optimizes document search and chat experiences, ensuring users receive the most relevant information. WebSockets efficiently stream chat messages, temporarily stored in a Redis database.

Docos

Landing page

This is Docos's landing page...

Personal Page

This is Docos's user page - home where he can see all his vaults

Chat window

From here the user can "talk with his documents"

Landing page

This is Docos's landing page...

Landing page

This is Docos's landing page...

Technical description

DocOS is one of the most complex projects I've developed, integrating multiple systems (pipelines):

Document Processing Pipeline

This pipeline analyzes uploaded documents, splitting them into semantically relevant chunks, and qualifying, summarizing, and extracting data. The documents are stored in three databases: PostgreSQL, Vectorized database (embeddings), and OpenSearch (or ElasticSearch). The pipeline also reviews existing documents for redundancy and conflicts.

Database Search Pipeline

This pipeline searches multiple databases using keywords and phrases, scoring each result. It employs "prompt engineering" to refine search terms for optimal database queries.

Multi-LLM Pipeline

The platform integrates multiple LLMs for various tasks. It supports local LLM deployments for privacy, with a default small local LLM for prompt engineering and large LLMs for chat-like experiences (e.g., GPT-4, Claude).

Prompt Engineering and WebSocket Pipeline

This layer optimizes document search and chat experiences, ensuring users receive the most relevant information. WebSockets efficiently stream chat messages, temporarily stored in a Redis database.

Swipeless

Demo

These are some of Swipeless's pages

Swipeless

Swipeless combats social isolation and smartphone addiction by promoting real-life experiences with new people. It groups users with 5-7 compatible individuals to foster genuine connections.

Skill Set & Technologies

FullStack
Payment processors
AI & Machine Learning
DevOps
RabbitMQ
Docker
Python/Django
Html/React
React Native
Postgresql
2023-2024
CTO/Lead developer

Technical description

Swipeless is a straightforward yet powerful platform leveraging AI for dynamic matchmaking using "Reinforcement Learning from Human Feedback".
Users sign up and complete a questionnaire covering various topics from political views to hobbies. The AI uses these responses to suggest compatible groups.
Human adjustments to these groups provide additional learning data for the AI.
The AI also selects the most suitable games to facilitate interactions during activities, helping to break the ice and spark meaningful conversations.

Technical description

Swipeless is a straightforward yet powerful platform leveraging AI for dynamic matchmaking using "Reinforcement Learning from Human Feedback".
Users sign up and complete a questionnaire covering various topics from political views to hobbies. The AI uses these responses to suggest compatible groups.
Human adjustments to these groups provide additional learning data for the AI.
The AI also selects the most suitable games to facilitate interactions during activities, helping to break the ice and spark meaningful conversations.

More projects coming soon